AD_classifiaction_project.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "midterm_project_2016250067.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "3NOKoAWa-K90",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4d86399f-1028-4949-b72c-2de7fe39f734"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GHBaj2bo-OlB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "03ba4ca2-2762-470c-9757-ab59334b1e2d"
      },
      "source": [
        "%cd /content/drive/My Drive/midterm"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/drive/My Drive/midterm\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WhZyyLn1V0LP"
      },
      "source": [
        "import pandas as pd\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "from torch.utils.data import Dataset\n",
        "from sklearn.preprocessing import StandardScaler, RobustScaler\n",
        "from sklearn.model_selection import KFold\n",
        "from sklearn.model_selection import StratifiedKFold\n",
        "import numpy as np\n",
        "from matplotlib import pyplot as plt\n",
        "from pytorchtools import EarlyStopping\n",
        "\n",
        "from collections import Counter\n",
        "from imblearn.over_sampling import SMOTE, RandomOverSampler\n",
        "from imblearn.under_sampling import TomekLinks, RandomUnderSampler\n",
        "from imblearn.combine import SMOTETomek\n",
        "from sklearn.decomposition import PCA\n",
        "from sklearn.neighbors import KNeighborsRegressor\n",
        "from sklearn.ensemble import GradientBoostingClassifier, AdaBoostClassifier\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import warnings\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IvDKwhJTmK2g"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from lightgbm import LGBMClassifier\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.metrics import accuracy_score, recall_score, precision_score\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "import lightgbm as lgb\n",
        "from sklearn.metrics import mean_squared_error, r2_score\n",
        "from sklearn.preprocessing import StandardScaler \n",
        "\n",
        "from scipy.optimize import minimize\n",
        "from sklearn.svm import LinearSVC\n",
        "from sklearn.ensemble import BaggingClassifier\n",
        "from sklearn.preprocessing import PolynomialFeatures\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "#Regression\n",
        "from sklearn.linear_model import LinearRegression, LogisticRegression\n",
        "import pickle\n",
        "from sklearn.externals import joblib"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ypDkeHhvgVv1"
      },
      "source": [
        "# Dataset 불러들임\n",
        "warnings.filterwarnings(action='ignore')\n",
        "\n",
        "filename = '/content/drive/My Drive/midterm/train_data.csv'\n",
        "data = pd.read_csv(filename)\n",
        "data = data.fillna(data.mean())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GsHtUyakcWRm"
      },
      "source": [
        "# Data analysis\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WQ3ER00lmZ40",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "134365b6-bc0b-4601-de3b-6f21b15b60a3"
      },
      "source": [
        "#pd.options.display.max_rows=999\n",
        "missing = data.isnull().sum()\n",
        "#ST123CV, ST64CV, ST123TA, ST64TA is the missing feature data\n",
        "\n",
        "count = data[\"DX_bl\"].value_counts() #class 별 개수 \n",
        "print(count, '\\n\\n')\n",
        "print('Proportion:', round(count[0] / count[2], 2),':', round(count[1] / count[2], 2), ': 1')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1    814\n",
            "0    485\n",
            "2    308\n",
            "Name: DX_bl, dtype: int64 \n",
            "\n",
            "\n",
            "Proportion: 1.57 : 2.64 : 1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-4Zhx4iE-YM0"
      },
      "source": [
        "**1.** First of all, it was examined whether missing features exist in the data of Column E-BV and Column BW-EN. As a result, there were 847 missing features among 1607 subjects in **ST123CV, ST64CV, ST123TA and ST64TA.** (one missing features occured in ST24CV, ST44CV, ST24TA and ST44TA.) So, the missing features were filled with the mean of the feaure(column).\n",
        "\n",
        "**2.** When comparing the number of data, the Mild cognitive impairment group was the largest with 814, the normal group was 485, and the AD group was 308. The MCI group, which is the median value between the normal group and the AD group, is the most common, and this data is imbalanced. The proportion between each data is **1.6: 2.6: 1**.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CGlKkQsoKmha"
      },
      "source": [
        "![22.JPG](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAeAB4AAD/4RD6RXhpZgAATU0AKgAAAAgABAE7AAIAAAAQAAAISodpAAQAAAABAAAIWpydAAEAAAAgAAAQ0uocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEd5ZW9uZ055ZW9uIEtpbQAABZADAAIAAAAUAAAQqJAEAAIAAAAUAAAQvJKRAAIAAAADNDkAAJKSAAIAAAADNDkAAOocAAcAAAgMAAAInAAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADIwMjA6MTA6MzEgMDA6NDA6MzkAMjAyMDoxMDozMSAwMDo0MDozOQAAAEcAeQBlAG8AbgBnAE4AeQBlAG8AbgAgAEsAaQBtAAAA/+ELImh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMjAtMTAtMzFUMDA6NDA6MzkuNDg1PC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPkd5ZW9uZ055ZW9uIEtpbTwvcmRmOmxpPjwvcmRmOlNlcT4NCgkJCTwvZGM6Y3JlYXRvcj48L3JkZjpEZXNjcmlwdGlvbj48L3JkZjpSREY+PC94OnhtcG1ldGE+DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgIDw/eHBhY2tldCBlbmQ9J3cnPz7/2wBDAAcFBQYFBAcGBQYIBwcIChELCgkJChUPEAwRGBUaGRgVGBcbHichGx0lHRcYIi4iJSgpKywrGiAvMy8qMicqKyr/2wBDAQcICAoJChQLCxQqHBgcKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKir/wAARCAFhAk8DASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwD6RooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiimo6SLujZXU91ORQA6iiigAoprukaF5GVFHVmOAKje7to7uO1e4iW4lUtHCXAdwOpA6kCgCaiiigAooppdFdUZlDNnaCeT9KAHUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeb+D/B+ha7pmo3Wp2Cy3Lalcp54dlcAPxgg8V6RXnPhrXbrw5Y39jP4b124na/nljMNk3luGbj5zgY96ANDwnqt5p8Gv6Xem51M6HNtgZF3zSxsCVX/AGmGMVdi8WajBq9haa5oD6dFqLmO3mF0spD4yFdQBt/M1lWmi+I7Twr4i1ONBBr+ruZkgjYHyVHCoD03BSefXFYttpYl1zw7PYaHrySwXaNfXeo+acnac8MSOv8AEAB0GeaAN3RdZ8Ry/EDW7aXTA8CtaiSNtQytohVvnUbcMWHzEDHTGTWlqvivVtPW9uovDE8unWTN5txJcrG7Kv3nSPBJXgkEkZHNV7eW50f4kas82mX01vqy2qw3NvAXjQoCrb2H3QN2eewrlNS0++vLPW4td0bXdR1qR5ltHiaT7LGhB2FSrBcDrtIJPTHNAHq1pdR3tlBdQZMU8ayJkYOGGR/OuHtvDekeIPiF4rOsWMd0YXtRGXJG0GAZxg+wrrvD8UkPhnTIpkaORLOJXRxgqQgyCOxrlF1Wfw9478STTaHrF5FetbGGWysmkRtsIB+bgdTigCfw4s/h/wAd3vhhJ5JtOeyF/aLK5doBv2MgJ5xnn/JqHR/EVhovgHTbvSNGMQvr1rW2sVucjzTI45kboDsJ9s4q34asdT1HxVe+KNYtHsPNtltLO0kILrEG3EtjoS3OPrWbpNrBa/DCwsPEWgaheo802+3htWaSI+c7KxAwy8dCPX3oA1Nd1fxCvgjVrh9FFjdRwyAGO/B2J5ZJlVgucqf4cA+9Q23inVNK+Hh1nV9KBNtawNH/AKZva5DBQWJ2/KeQec9ap6NYavceEfE9l5eofYp4JItLi1L/AF4DREEHPIG4jGai1OW71f4QXWnQaPqcV5bWlvCYZrVlaRgVB2DqwG08igC5revPf+BdWvdf8MyLpipE8MMt3ta6UuME7RmMjCnHPX61H4mupNL+Ivh42NjJeSCzniit4mAyeAMseAAOST6Vr/EO1uL3wBqlvZQS3E8iIEiiQuzfOp4A5NUPExv7Px7o2qWml3V/b21rMs/2dMlQxUcdi3fbnJANAGpoXiebUtYu9H1XTW03UbaNZvK84SrJGeNysAO/HSorbxf9o+Hh8UfYduIHm+zebn7rEY3Y9uuKp6Gl5rHj+78RS6fdWFnHp62MC3cZjklO/ezbTyADxXNxPrFp8M7nwn/wj2pSahDDNEZFh/csu5m3K/8AESOAo5JoA1vE/iDWmufCtxpVmwhvZI5dgvfL85mQnyWG3oBg7jx7VrXVxC/i3wydV0dU1SeK5Mci3RYWu1BuHAAfIbGTjFZms2l9b6B4Pu0sLq4/s2WB7mCCItKg8rB+TqcHqK0L37RqfjLwnqMNjdxwLFeGXzYSph3IoUP/AHSccA0APl8W391qV9a+HdCbU49Pk8m4ne6WBfMHVFyDuI79KjuPH9rH4LXxDbWcsoE6wS2rNteN921lPByR+vHSs/Rrq78GXOr6fd6Lql7FPfSXdrcWVuZlkV8fKxH3WGO+BVC50DVYfh3Ks9k5vr7V1vntYELmENKpxhfQDJ9KAOktvF16mt2NjrWhSabHqO4WszXCyEsBna6gfIce5rqK5PxfZXV14k8LS21tLNHBfl5njQsI129WI6D611lABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUARXV1DZWc11dP5cEEbSSORnaqjJPHsKyLTxVazMwvbW603FqbxPtap+8hH3mGxm+7kZBweRxWtdxpLZTxywfaUeNlaHj94COV5IHPTkgVxK+Ep9QmvPKs7vTLVdLmsbaK+vmuG3ygAsB5kgRQFAwDk+nAoA2D42sYIfN1CzvrFXtvtNuJ41JuUyBhArH5sunytg/MOOuNLTdbTUL24spbS5sby3VZGgudm5kbOHUozKRkEcHgjmuS1jQ9b8RWtizaYbGXSrdWVJ5oz9pm3xttUqWwuIjy2OWXjrjodMtby68UXOtXlnJYJ9lS0igmdGkbDM7OdjMAPmAAyTwTxxQBvUUUx0ZyNsrp7KBz+YoAfRUXkv/z8y/kv+FHkv/z8y/kv+FAEtFReS/8Az8y/kv8AhR5L/wDPzL+S/wCFAEtFReS//PzL+S/4UeS//PzL+S/4UAS0VF5L/wDPzL+S/wCFHkv/AM/Mv5L/AIUAS0VF5L/8/Mv5L/hR5L/8/Mv5L/hQBLRUXkv/AM/Mv5L/AIUeS/8Az8y/kv8AhQBLRUXkv/z8y/kv+FHkv/z8y/kv+FAEtFReS/8Az8y/kv8AhR5L/wDPzL+S/wCFAEtFReS//PzL+S/4UeS//PzL+S/4UAS0VF5L/wDPzL+S/wCFHkv/AM/Mv5L/AIUAS0VF5L/8/Mv5L/hR5L/8/Mv5L/hQBLRUXkv/AM/Mv5L/AIUeS/8Az8y/kv8AhQBLRUXkv/z8y/kv+FHkv/z8y/kv+FAEtFReS/8Az8y/kv8AhR5L/wDPzL+S/wCFAEtFReS//PzL+S/4UeS//PzL+S/4UAS0VF5L/wDPzL+S/wCFHkv/AM/Mv5L/AIUAS0VF5L/8/Mv5L/hR5L/8/Mv5L/hQBLRWbqeoW2j26TX97OiySCKNY4fMeRz0VURCzHAJwAeAT2qlJ4o0aHR21SfV5IrZXeNvMi2yB1zuTyym/cACSMZwM9OaAN+is+4vba00l9SuL+SOzSLzTKVH3cZ6bc59uvan2FzFqmnW99Y3ssttcxiSJ9gXcpGQcFQR+NAF2iovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf8A5+ZfyX/CjyX/AOfmX8l/woAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWikUFVALFj6nvS0AFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBzHjeOWW30sJBdGFL5ZJbuzheWa0wjYdETJJJO0khgAxypBrFiM9p8NtdtZLLUJDdPdxWb/YZmuLvehxJKgUsrFiwywUYC4CggV6DRQBxN3batrOl6M+nWkUtlZxb5rS/aazkedFAQ7TESVU5IGAC2054rV8BQ3tt4D0i31O1+yzw2saeUWYsAFGNwKgq3qvOPU10NFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAEN7JJDYXEkBjWRImZDMcICBxuPp61wumeKNSihvbie7nuhbaM968V5HEm6VehhMYG+Lg/OCw5TDcmu6vLSG/sZ7O6XfDcRtFIucZVhgj8jWPB4Stg6nUL691NY7Z7WNLpkASNwAwGxVySABk5NAHO6zrmueHYLAHU/t0mq2wUNPBGotpi8S71CgZX96flbP3Rz1z0OmXN7aeKrrRry9kv4vskd3DNMiLImWZGQ7FUEcAg4B5PWoz4KsprYw6he318FtxbwNPIga2UEHKFVX5sqh3Nk/KOeudLTdGSwup7uW6uL67nCo9xc7NwRc7UARVUAEseBySc54oA0aY80URAkkRCem5gKfRQBF9rt/wDnvF/32KPtdv8A894v++xUtFAEX2u3/wCe8X/fYo+12/8Az3i/77FS0UARfa7f/nvF/wB9ij7Xb/8APeL/AL7FS0UARfa7f/nvF/32KPtdv/z3i/77FS0UARfa7f8A57xf99ij7Xb/APPeL/vsVLRQBF9rt/8AnvF/32KPtdv/AM94v++xUtFAEX2u3/57xf8AfYo+12//AD3i/wC+xUtFAEX2u3/57xf99ij7Xb/894v++xUtFAEX2u3/AOe8X/fYo+12/wDz3i/77FS0UARfa7f/AJ7xf99ij7Xb/wDPeL/vsVLRQBF9rt/+e8X/AH2KPtdv/wA94v8AvsVLRQBF9rt/+e8X/fYo+12//PeL/vsVLRQBF9rt/wDnvF/32KPtdv8A894v++xUtFAEX2u3/wCe8X/fYo+12/8Az3i/77FS0UARfa7f/nvF/wB9ij7Xb/8APeL/AL7FS0UARfa7f/nvF/32KPtdv/z3i/77FS0UARfa7f8A57xf99iuQ8WeM7zRNVittPS1miaASFnBYglmGOGHoK7SvL/iX/yMlv8A9ei/+hvXHjJyhSvF2ZjXk4wug/4WXrH/AD7WP/ft/wD4qj/hZesf8+1j/wB+3/8Aiq4+ivF+tV/5jh9tU7nYf8LL1j/n2sf+/b//ABVH/Cy9Y/59rH/v2/8A8VXH0UfWq/8AMHtqnc9f8J+I5Nb0qW51A28MqzmMKh2ggKpzyT6mtz7Xb/8APeL/AL7Fcp8NP+RbuP8Ar7b/ANASuwr38PJypRb3PQptuCbIvtdv/wA94v8AvsUfa7f/AJ7xf99ipaK3NCL7Xb/894v++xR9rt/+e8X/AH2KlooAi+12/wDz3i/77FH2u3/57xf99ipaKAIvtdv/AM94v++xR9rt/wDnvF/32KlooAi+12//AD3i/wC+xR9rt/8AnvF/32KlooAi+12//PeL/vsUfa7f/nvF/wB9ipaKAIvtdv8A894v++xR9rt/+e8X/fYqWigCL7Xb/wDPeL/vsUfa7f8A57xf99ipaKAIvtdv/wA94v8AvsUfa7f/AJ7xf99ipaKAIvtdv/z3i/77FH2u3/57xf8AfYqWigCL7Xb/APPeL/vsUfa7f/nvF/32KlooAi+12/8Az3i/77FH2u3/AOe8X/fYqWigCL7Xb/8APeL/AL7FH2u3/wCe8X/fYqWigBFYOoZSGB6EHrS0UUAFFMlljgheaeRY4o1LO7nCqBySSegqlZ+ING1G4EGn6tY3UxBIjguUdsDvgHNAGhRVK21rSr0Tmz1OzuBbjdMYrhW8oc8tg8dD19KksdTsdUiaXTb23vI1baz28qyAH0JBPNAFmiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK8v8AiX/yMlv/ANei/wDob16hXl/xL/5GS3/69F/9DeuDH/wDnxH8M4+iiivnzzgooooA9Q+Gn/It3H/X23/oCV2Fcf8ADT/kW7j/AK+2/wDQErsK+mwv8GJ6lL+GgooorpNQooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK4mRJ9ftPEutWgeTdZz6fpgTqyqp3MPXdIMD/cFdtRQB5V4jmstWsNKbQTHcwWGnqNQFsNwig8yE+W4HsjnYeQFbiuu0m4t9S8dX+oaRNHcWP2GGGaeBw0ckwdyBkcMyowz6BgK6eigApjiUkeW6KO+5Cf6in0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FUbbj/nrF/36P8A8VUtFAEW24/56xf9+j/8VRtuP+esX/fo/wDxVS0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FUbbj/nrF/36P8A8VUtFAEW24/56xf9+j/8VRtuP+esX/fo/wDxVS0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FUbbj/nrF/36P8A8VUtFAEW24/56xf9+j/8VRtuP+esX/fo/wDxVS0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FUbbj/nrF/36P8A8VUtFAEW24/56xf9+j/8VRtuP+esX/fo/wDxVS0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FUbbj/nrF/36P8A8VUtFAEW24/56xf9+j/8VRtuP+esX/fo/wDxVS0UARbbj/nrF/36P/xVG24/56xf9+j/APFVLRQBFtuP+esX/fo//FV5l8SA48RweYysfsi8quP439zXqVeX/Ev/AJGS3/69F/8AQ3rgx/8AAOfEfwzj6KKK+fPOCiiigD034biU+HJ/LdFH2tuGQn+BPcV1224/56xf9+j/APFVynw0/wCRbuP+vtv/AEBK7CvpsL/BiepS/hoi23H/AD1i/wC/R/8AiqNtx/z1i/79H/4qpaK6TUi23H/PWL/v0f8A4qjbcf8APWL/AL9H/wCKqWigCLbcf89Yv+/R/wDiqNtx/wA9Yv8Av0f/AIqpaKAIttx/z1i/79H/AOKo23H/AD1i/wC/R/8AiqlooAi23H/PWL/v0f8A4qjbcf8APWL/AL9H/wCKqWigCLbcf89Yv+/R/wDiqNtx/wA9Yv8Av0f/AIqpaKAIttx/z1i/79H/AOKo23H/AD1i/wC/R/8AiqlooAi23H/PWL/v0f8A4qjbcf8APWL/AL9H/wCKqWigCLbcf89Yv+/R/wDiqNtx/wA9Yv8Av0f/AIqpaKAIttx/z1i/79H/AOKo23H/AD1i/wC/R/8AiqlooAi23H/PWL/v0f8A4qjbcf8APWL/AL9H/wCKqWigCLbcf89Yv+/R/wDiqNtx/wA9Yv8Av0f/AIqpaKAIttx/z1i/79H/AOKo23H/AD1i/wC/R/8AiqlooARdwUbiCe5AxS0UUAFFR3E8VrbS3Fw+yKJC7sf4VAyT+Vc5pfjOHULO7vn+wLaW1s9yyQX4muUQcjfGFwpxnPzHB4560AdPRXJXPjW40yGJ9Y0gQNd23nWaQ3PmGVtyL5TZVQjZkT1HXnjnW03WbifV7nS9Ts47S8hiSdRDOZkkjYkZDFVIIZSCCPTrQBr0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXl/wAS/wDkZLf/AK9F/wDQ3r1CvL/iX/yMlv8A9ei/+hvXBj/4Bz4j+GcfRRRXz55wUUUUAeofDT/kW7j/AK+2/wDQErsK4/4af8i3cf8AX23/AKAldhX02F/gxPUpfw0FFFFdJqFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAFfULRdQ026s3bYtxC8RYDOAwIz+tcsnhTUtRhhh1lrO3W20uXT45LR2dpfMUKXYMq7QAuQoJ5PXjnsaKAOK1HwprOu2tsNSksbabTrcLa+RI8iyzBkbe+VXav7oDaN33jzwM7em6bfNr1zrGrLbwzPAlrFBbStIqopLFizKuWJbpjgKOua2qKACmPDFKQZI0cjpuUGn0UARfZLf8A54Rf98Cj7Jb/APPCL/vgVLRQBF9kt/8AnhF/3wKPslv/AM8Iv++BUtFAEX2S3/54Rf8AfAo+yW//ADwi/wC+BUtFAEX2S3/54Rf98Cj7Jb/88Iv++BUtFAEX2S3/AOeEX/fAo+yW/wDzwi/74FS0UARfZLf/AJ4Rf98Cj7Jb/wDPCL/vgVLRQBF9kt/+eEX/AHwKPslv/wA8Iv8AvgVmeJfE1j4W0wXuoLNKGbakUChpG4ySASBgAEk5/pUGreLrbSbm8i+wXt2unwLcXstuse23jbcQTvdS3CMcKG6UAbX2S3/54Rf98Cj7Jb/88Iv++BWR/wAJOJdYm0/T9Jv74wJE8s0JhVEEgyp+eRWPAPQVuUARfZLf/nhF/wB8Cj7Jb/8APCL/AL4FS0UARfZLf/nhF/3wKPslv/zwi/74FS0UARfZLf8A54Rf98Cj7Jb/APPCL/vgVLRQBF9kt/8AnhF/3wKPslv/AM8Iv++BUtFAEX2S3/54Rf8AfAo+yW//ADwi/wC+BUtFAEX2S3/54Rf98Cj7Jb/88Iv++BUtFAEX2S3/AOeEX/fAo+yW/wDzwi/74FS0UARfZLf/AJ4Rf98Cj7Jb/wDPCL/vgVLRQBF9kt/+eEX/AHwK8y+JEaR+I4BGioPsinCjH8b16lXl/wAS/wDkZLf/AK9F/wDQ3rgx/wDAOfEfwzj6KKK+fPOCiiigD034bwRSeHJzJEjn7Wwyyg/wJXXfZLf/AJ4Rf98CuU+Gn/It3H/X23/oCV2FfTYX+DE9Sl/DRF9kt/8AnhF/3wKPslv/AM8Iv++BUtFdJqRfZLf/AJ4Rf98Cj7Jb/wDPCL/vgVLRQBF9kt/+eEX/AHwKPslv/wA8Iv8AvgVLRQBF9kt/+eEX/fAo+yW//PCL/vgVLRQBF9kt/wDnhF/3wKPslv8A88Iv++BUtFAEX2S3/wCeEX/fAo+yW/8Azwi/74FS0UARfZLf/nhF/wB8Cj7Jb/8APCL/AL4FS0UARfZLf/nhF/3wKPslv/zwi/74FS0UARfZLf8A54Rf98Cj7Jb/APPCL/vgVLRQBF9kt/8AnhF/3wKPslv/AM8Iv++BUtFAEX2S3/54Rf8AfAo+yW//ADwi/wC+BUtFAEX2S3/54Rf98Cj7Jb/88Iv++BUtFAEX2S3/AOeEX/fAo+yW/wDzwi/74FS0UAIqhFCqAoHQAdKWiigAoqK7jklsp44H2SvGyo2SNrEcHIrh4bq/8NWN9a3yTNqsOjSXcVz/AGlNdxTNGMMSkmAh3FTgAjBIz2IB3tFeba/cX3h630tdO1S+Y6paiO4ee4aUqfMhXzk3EhGxI3TA5HHArptMEmneNLvSobi5ms2sY7oLczvM0Tl2Q4ZyTghRwSeQcY5oA6OiimO7IRtid/dSOPzNAD6Ki85/+faX81/xo85/+faX81/xoAloqLzn/wCfaX81/wAaPOf/AJ9pfzX/ABoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf/AJ9pfzX/ABo85/8An2l/Nf8AGgCWiovOf/n2l/Nf8aPOf/n2l/Nf8aAJaKi85/8An2l/Nf8AGjzn/wCfaX81/wAaAON8beFNZ1dNTu9Nu7ad57IWsFpNbEsi5ywSTzVVSxAJJU/dA7VU8R+H9Y1DUbyS802TU5vskaWFzYNDBEkgJYiZJH3MvmBW2t5igdBnOe985/8An2l/Nf8AGjzn/wCfaX81/wAaAOK8RaHeaxC8H/CORf2yRCItfiaFREw2EybsiVSrbsKFOcDnBOO6qLzn/wCfaX81/wAaPOf/AJ9pfzX/ABoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf/AJ9pfzX/ABo85/8An2l/Nf8AGgCWiovOf/n2l/Nf8aPOf/n2l/Nf8aAJaKi85/8An2l/Nf8AGjzn/wCfaX81/wAaAJaKi85/+faX81/xo85/+faX81/xoAloqLzn/wCfaX81/wAaPOf/AJ9pfzX/ABoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf/AJ9pfzX/ABo85/8An2l/Nf8AGgCWvL/iX/yMlv8A9ei/+hvXpfnP/wA+0v5r/jXmXxIYt4jgLIyf6IvDY/vv6GuDH/wDnxH8M5GiiivnzzgooooA9Q+Gn/It3H/X23/oCV2FcZ8N5GXw5OFid/8AS25Uj+4nqa67zn/59pfzX/GvpsL/AAYnqUv4aJaKi85/+faX81/xo85/+faX81/xrpNSWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWiovOf8A59pfzX/Gjzn/AOfaX81/xoAloqLzn/59pfzX/Gjzn/59pfzX/GgCWikUkqCVKn0PaloAjuLeK7tZba5QSQzIY5EPRlIwR+VUbTQNOs5JpEhkmknj8qR7qeS4Zk5+TMjMQvJ46VpUUAY0PhLRYbeaD7G0sc0At2WeeSXbEOQi72OwZwcLjkD0GLmnaRZ6WZmtEk8ycgyyzTPNI+BgAu5LYHYZwMn1NXaKACiiigAooooAKKKKACiiigAooooAKKKKACvK9X8Za9a65fW8F/siiuJERfJjOFDEAcrXqleH6/8A8jJqf/X3L/6Ga83MJyhGPK7HLiJNJWZof8Jz4i/6CP8A5Aj/APiaP+E58Rf9BH/yBH/8TXP0V5Ht6v8AM/vZx+0n3Z0H/Cc+Iv8AoI/+QI//AImj/hOfEX/QR/8AIEf/AMTXP0Ue3q/zP72HtJ92e6aRPJdaHY3E7b5ZbeN3bAGWKgk8VcrP0D/kW9M/69Iv/QBWhX00NYo9WOyCiiirGFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeX/Ev/AJGS3/69F/8AQ3r1CvL/AIl/8jJb/wDXov8A6G9cGP8A4Bz4j+GcfRRRXz55wUUUUAeofDT/AJFu4/6+2/8AQErsK4/4af8AIt3H/X23/oCV2FfTYX+DE9Sl/DQUUUV0moUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFZ8Ou6bPqV5YRXaG4sUD3CkECMHuWIx2PfjvVjUBdtptwNMMS3hjYQNMTsV8cE4B4B9q4KLQNbjudWsP7OhT7RoJt/tCXTyebMzTHcXaNQXZnyw4xkHJzgAHVw+LtEmhllF4Yo4rf7SWngkiDRdN671G9eRyueo9RVzTtZstVeeOzeTzbcqJYpoHhdNwyCVcA4ODg4wcH0rhNftbzxDaaZJp2mXqnSbYSXCTWzxMzeZCTEgYDecRv93I4HPIrp9MD6j4yu9Xit7iG0WyjtFa5geFpXDs5wrgNtAYDJHJJx0NAHRUUUx0ZyNsrp7KBz+YoAfRUXkv/AM/Mv5L/AIUeS/8Az8y/kv8AhQBLRUXkv/z8y/kv+FHkv/z8y/kv+FAEtFReS/8Az8y/kv8AhR5L/wDPzL+S/wCFAEtFReS//PzL+S/4UeS//PzL+S/4UAS0VF5L/wDPzL+S/wCFHkv/AM/Mv5L/AIUAS14fr/8AyMmp/wDX3L/6Ga9r8l/+fmX8l/wrxPXhjxHqQJJ/0uXk9/nNeVmXwROTFbIoUUUV4pwhRRRQB7hoH/It6Z/16Rf+gCtCsrQYnPhzTSLiQf6JFwAvHyD2rQ8l/wDn5l/Jf8K+sp/Aj147IloqLyX/AOfmX8l/wo8l/wDn5l/Jf8KsoloqLyX/AOfmX8l/wo8l/wDn5l/Jf8KAJaKi8l/+fmX8l/wo8l/+fmX8l/woAloqLyX/AOfmX8l/wo8l/wDn5l/Jf8KAJaKi8l/+fmX8l/wo8l/+fmX8l/woAloqLyX/AOfmX8l/wo8l/wDn5l/Jf8KAJaKi8l/+fmX8l/wo8l/+fmX8l/woAlry/wCJf/IyW/8A16L/AOhvXpfkv/z8y/kv+FeZfEhSniOAM7P/AKIvLY/vv6CuDH/wDnxH8M5GiiivnzzgooooA9Q+Gn/It3H/AF9t/wCgJXYVxnw3jZ/Dk5WV0/0tuFA/uJ6iuu8l/wDn5l/Jf8K+mwv8GJ6lL+GiWiovJf8A5+ZfyX/CjyX/AOfmX8l/wrpNSWiovJf/AJ+ZfyX/AAo8l/8An5l/Jf8ACgCWiovJf/n5l/Jf8KPJf/n5l/Jf8KAJaKi8l/8An5l/Jf8ACjyX/wCfmX8l/wAKAJaKi8l/+fmX8l/wo8l/+fmX8l/woAloqLyX/wCfmX8l/wAKPJf/AJ+ZfyX/AAoAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWiovJf/AJ+ZfyX/AAo8l/8An5l/Jf8ACgCWiovJf/n5l/Jf8KPJf/n5l/Jf8KAJaKi8l/8An5l/Jf8ACjyX/wCfmX8l/wAKAJaKi8l/+fmX8l/wo8l/+fmX8l/woAloqLyX/wCfmX8l/wAKPJf/AJ+ZfyX/AAoAloqLyX/5+ZfyX/CjyX/5+ZfyX/CgCWikUFVALFj6nvS0AFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV4fr/wDyMmp/9fcv/oZr3CvD9f8A+Rk1P/r7l/8AQzXlZl8ETkxWyM+iiivFOEKKKKAPcNA/5FvTP+vSL/0AVoVn6B/yLemf9ekX/oArQr6yn8CPXjsgoooqygooooAKKKKACiiigAooooAKKKKACiiigAry/wCJf/IyW/8A16L/AOhvXqFeX/Ev/kZLf/r0X/0N64Mf/AOfEfwzj6KKK+fPOCiiigD1D4af8i3cf9fbf+gJXYVx/wANP+RbuP8Ar7b/ANASuwr6bC/wYnqUv4aCiiiuk1CiiigAooooAKKKKACiiigAooooAKKKKACiiigCvqH2f+zbn7bIsdsYmErvKYwqkckuMFeO/avMNShbw8viOCGzt9Fmm0qGUW+lSZhWISlHkLbUIkIYjIQYC53E5x6rJGk0TxTIskbqVZGGQwPUEdxVOx0TStMhlh03TLOzim/1qW9ukayduQBz170Aed+I7PTNFv8AULDRx9gsLnR4hNHYoNhla4VIiw3KPnBZSSRlc5Nb3gq2/szxDrenz2tpYXRS3nNnpw/0VEKsoZeh3kqd2VXgL1HJ6S38P6NaWU9naaTYwWtx/roIrZFSXt8ygYP41Lp+k6dpETxaVYWtjG7bmS2hWMMfUhQMmgC3RRRQAUUUUAFFFFABRRRQA2QOYmETKrkHazLuAPYkZGfpkVxVv4gvrCTWC19cX8VnpRu1Go2q20rSruztQIjGM4GSQcHgN1rs7iEXFtLCXkjEiFC8bFWXIxkEdD71lQ+GYPPmm1G8u9TkmtWtC10UG2JjllAjVRzgZJyeBzQBzGs65rnh2CwB1P7dJqtsFDTwRqLaYvEu9QoGV/en5Wz90c9c9Dplze2niq60a8vZL+L7JHdwzTIiyJlmRkOxVBHAIOAeT1qM+CrKa2MOoXt9fBbcW8DTyIGtlBByhVV+bKodzZPyjnrnS03RksLqe7luri+u5wqPcXOzcEXO1AEVVABLHgcknOeKANGmPNFEQJJEQnpuYCn0UARfa7f/AJ7xf99ij7Xb/wDPeL/vsVLRQBF9rt/+e8X/AH2KPtdv/wA94v8AvsVLRQBF9rt/+e8X/fYo+12//PeL/vsVLRQBF9rt/wDnvF/32KPtdv8A894v++xUtFAEX2u3/wCe8X/fYo+12/8Az3i/77FS0UARfa7f/nvF/wB9ivE9eIbxHqRUgg3cpBHf5zXuNeH6/wD8jJqf/X3L/wChmvKzL4InJitkZ9FFFeKcIUUUUAe16DcwL4c00NNGCLSIEFxx8grQ+12//PeL/vsVU0D/AJFvTP8Ar0i/9AFaFfWU/gR68dkRfa7f/nvF/wB9ij7Xb/8APeL/AL7FS0VZRF9rt/8AnvF/32KPtdv/AM94v++xUtFAEX2u3/57xf8AfYo+12//AD3i/wC+xUtFAEX2u3/57xf99ij7Xb/894v++xUtFAEX2u3/AOe8X/fYo+12/wDz3i/77FS0UARfa7f/AJ7xf99ij7Xb/wDPeL/vsVLRQBF9rt/+e8X/AH2KPtdv/wA94v8AvsVLRQBF9rt/+e8X/fYrzL4kSJJ4jgMbq4+yKMqc/wAb16lXl/xL/wCRkt/+vRf/AEN64Mf/AADnxH8M4+iiivnzzgooooA9N+G88UfhycSSoh+1scMwH8CV132u3/57xf8AfYrlPhp/yLdx/wBfbf8AoCV2FfTYX+DE9Sl/DRF9rt/+e8X/AH2KPtdv/wA94v8AvsVLRXSakX2u3/57xf8AfYo+12//AD3i/wC+xUtFAEX2u3/57xf99ij7Xb/894v++xUtFAEX2u3/AOe8X/fYo+12/wDz3i/77FS0UARfa7f/AJ7xf99ij7Xb/wDPeL/vsVLRQBF9rt/+e8X/AH2KPtdv/wA94v8AvsVLRQBF9rt/+e8X/fYo+12//PeL/vsVLRQBF9rt/wDnvF/32KqanrdnpWnS3kz+akeMpEQWOSBwMj1rQrn/ABz/AMiXff8AbP8A9GLWdWTjTlJdEyZu0W0Z/wDwsvR/+fa+/wC/af8AxVH/AAsvR/8An2vv+/af/FV5fRXh/X65wfWKh6h/wsvR/wDn2vv+/af/ABVH/Cy9H/59r7/v2n/xVeX0UfX64fWKh7Bo3jPTdbvHtrdJ4WWMyFpwqqQCBjhjzzW39rt/+e8X/fYrzT4af8jJcf8AXo3/AKGleoV62Fqyq0+aW52UZOcbsi+12/8Az3i/77FH2u3/AOe8X/fYqWiuo1Ivtdv/AM94v++xR9rt/wDnvF/32KlooARWDqGUhgehB60tFFABRRVSw1bTtVVzpl/a3gjwHNvMsm3PTOCcdKALdFUrbWtKvRObPU7O4FuN0xiuFbyhzy2Dx0PX0qSx1Ox1SJpdNvbe8jVtrPbyrIAfQkE80AWaKKKACiiigAooooAKKKKACiiigAooooAK8P1//kZNT/6+5f8A0M17hXh+v/8AIyan/wBfcv8A6Ga8rMvgicmK2Rn0UUV4pwhRRRQB7hoH/It6Z/16Rf8AoArQrP0D/kW9M/69Iv8A0AVoV9ZT+BHrx2QUUUVZQUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV5f8S/+Rkt/+vRf/Q3r1CvL/iX/AMjJb/8AXov/AKG9cGP/AIBz4j+GcfRRRXz55wUUUUAeofDT/kW7j/r7b/0BK7CuP+Gn/It3H/X23/oCV2FfTYX+DE9Sl/DQUUUV0moUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVz/jn/kS77/tn/wCjFroK5/xz/wAiXff9s/8A0YtY1/4UvR/kRU+B+h4/RRRXyx5IUUUUAdh8NP8AkZLj/r0b/wBDSvUK8v8Ahp/yMlx/16N/6GleoV9BgP4J6OH/AIYUUUV3nQFFFFABRRRQBFdtAlnM15s+zrGxl8wZXZjnPtiuCjv7TxJDr1/pFzbfaJNJks9Ps7eVTMY1DEOyA5XLMAq8EDrycD0KigDyrxHNZatYaU2gmO5gsNPUagLYbhFB5kJ8twPZHOw8gK3FddpNxb6l46v9Q0iaO4sfsMMM08Dho5Jg7kDI4ZlRhn0DAV09FABTHMoI8tEYd9zkf0NPooAi3XH/ADyi/wC/p/8AiaN1x/zyi/7+n/4mpaKAIt1x/wA8ov8Av6f/AImjdcf88ov+/p/+JqWigCLdcf8APKL/AL+n/wCJo3XH/PKL/v6f/ialooAi3XH/ADyi/wC/p/8AiaN1x/zyi/7+n/4mpaKAIt1x/wA8ov8Av6f/AImjdcf88ov+/p/+JqWigCLdcf8APKL/AL+n/wCJrxPXs/8ACR6luAB+1y5AOf4zXuNeH6//AMjJqf8A19y/+hmvKzL4InJitkZ9FFFeKcIUUUUAe16C0/8Awjmm7Y4yPskWCZCP4B/s1obrj/nlF/39P/xNVNA/5FvTP+vSL/0AVoV9ZT+BHrx2RFuuP+eUX/f0/wDxNG64/wCeUX/f0/8AxNS0VZRFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATRuuP+eUX/f0//E1LRQBFuuP+eUX/AH9P/wATXmXxILnxHB5iqp+yLwrZ/jf2FepV5f8AEv8A5GS3/wCvRf8A0N64Mf8AwDnxH8M4+iiivnzzgooooA9N+G5lHhyfy0Rh9rblnI/gT2NdduuP+eUX/f0//E1ynw0/5Fu4/wCvtv8A0BK7CvpsL/BiepS/hoi3XH/PKL/v6f8A4mjdcf8APKL/AL+n/wCJqWiuk1It1x/zyi/7+n/4mjdcf88ov+/p/wDialooAi3XH/PKL/v6f/iaN1x/zyi/7+n/AOJqWigCLdcf88ov+/p/+Jo3XH/PKL/v6f8A4mpaKAIt1x/zyi/7+n/4mjdcf88ov+/p/wDialooAi3XH/PKL/v6f/iaN1x/zyi/7+n/AOJqWigCLdcf88ov+/p/+Jo3XH/PKL/v6f8A4mpaKAIt1x/zyi/7+n/4msDxu0x8G32+OML+7yQ5J/1i+1dJXP8Ajn/kS77/ALZ/+jFrGv8Awpej/Iip8D9Dx+iiivljyQooooA674blx4jn8tVY/ZG4Zsfxp7GvTd1x/wA8ov8Av6f/AImvNPhp/wAjJcf9ejf+hpXqFfQYD+Cejh/4ZFuuP+eUX/f0/wDxNG64/wCeUX/f0/8AxNS0V3nQRbrj/nlF/wB/T/8AE0brj/nlF/39P/xNS0UAIu4qNwAPcA5paKKACikdiqMwUuQMhVxk+3PFYVj4m8601m41CyayXSXbzE8wSMVEYkyccA4bBAJGR1PWgDeorkrnxrcaZDE+saQIGu7bzrNIbnzDK25F8psqoRsyJ6jrzxzrabrNxPq9zpep2cdpeQxJOohnMySRsSMhiqkEMpBBHp1oA16KKKACiiigAooooAKKKKACiiigAooooAK8P1//AJGTU/8Ar7l/9DNe4V4fr/8AyMmp/wDX3L/6Ga8rMvgicmK2Rn0UUV4pwhRRRQB7hoH/ACLemf8AXpF/6AK0Kz9A/wCRb0z/AK9Iv/QBWhX1lP4EevHZBRRRVlBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXl/xL/wCRkt/+vRf/AEN69Qry/wCJf/IyW/8A16L/AOhvXBj/AOAc+I/hnH0UUV8+ecFFFFAHqHw0/wCRbuP+vtv/AEBK7CuP+Gn/ACLdx/19t/6AldhX02F/gxPUpfw0FFFFdJqFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFc/45/wCRLvv+2f8A6MWugrn/ABz/AMiXff8AbP8A9GLWNf8AhS9H+RFT4H6Hj9FFFfLHkhRRRQB2Hw0/5GS4/wCvRv8A0NK9Qry/4af8jJcf9ejf+hpXqFfQYD+Cejh/4YUUUV3nQFFFFABRRRQAjlgjGMBmx8oY4BP17VytpoOrTr4gt9UisoINaDkvb3LytEWiWPG0xqD0JzkemO9dXRQBxWo+FNZ121thqUljbTadbhbXyJHkWWYMjb3yq7V/dAbRu+8eeBnb03Tb5teudY1ZbeGZ4EtYoLaVpFVFJYsWZVyxLdMcBR1zW1RQAUx4lkILFxj+65X+Rp9FAEX2ZP70v/f1v8aPsyf3pf8Av63+NS0UARfZk/vS/wDf1v8AGj7Mn96X/v63+NS0UARfZk/vS/8Af1v8aPsyf3pf+/rf41LRQBF9mT+9L/39b/Gj7Mn96X/v63+NS0UARfZk/vS/9/W/xo+zJ/el/wC/rf41LRQBF9mT+9L/AN/W/wAa8T14bfEepAZ4u5Ryc/xmvca8P1//AJGTU/8Ar7l/9DNeVmXwROTFbIz6KKK8U4QooooA9r0G3RvDmmkmTm0iPErD+Ae9aH2ZP70v/f1v8aqaB/yLemf9ekX/AKAK0K+sp/Aj147Ii+zJ/el/7+t/jR9mT+9L/wB/W/xqWirKIvsyf3pf+/rf40fZk/vS/wDf1v8AGpaKAIvsyf3pf+/rf40fZk/vS/8Af1v8alooAi+zJ/el/wC/rf40fZk/vS/9/W/xqWigCL7Mn96X/v63+NH2ZP70v/f1v8alooAi+zJ/el/7+t/jR9mT+9L/AN/W/wAalooAi+zJ/el/7+t/jR9mT+9L/wB/W/xqWigCL7Mn96X/AL+t/jXmXxIQR+I4ApY/6Ip+Zif439a9Sry/4l/8jJb/APXov/ob1wY/+Ac+I/hnH0UUV8+ecFFFFAHpvw3hWTw5OWLj/S2HyuR/Anoa677Mn96X/v63+Ncp8NP+RbuP+vtv/QErsK+mwv8ABiepS/hoi+zJ/el/7+t/jR9mT+9L/wB/W/xqWiuk1Ivsyf3pf+/rf40fZk/vS/8Af1v8alooAi+zJ/el/wC/rf40fZk/vS/9/W/xqWigCL7Mn96X/v63+NH2ZP70v/f1v8alooAi+zJ/el/7+t/jR9mT+9L/AN/W/wAalooAi+zJ/el/7+t/jR9mT+9L/wB/W/xqWigCL7Mn96X/AL+t/jR9mT+9L/39b/GpaKAIvsyf3pf+/rf41geN4FTwbfMDJkeX1kYj/WL2JrpK5/xz/wAiXff9s/8A0YtY1/4UvR/kRU+B+h4/RRRXyx5IUUUUAdd8N0EniOcMWH+iMflYj+NPSvTfsyf3pf8Av63+NeafDT/kZLj/AK9G/wDQ0r1CvoMB/BPRw/8ADIvsyf3pf+/rf40fZk/vS/8Af1v8alorvOgi+zJ/el/7+t/jR9mT+9L/AN/W/wAalooARVCqAM4Hqc0tFFABRRRQAUVi+L7u4sfCt3NZytDMTHGJV6oHkVCR7gMTVTTEk0zxpPpcN1dT2b6elztup3maOTeU4ZySAQOmcZHFAHS0UUx3ZCNsTv7qRx+ZoAfRUXnP/wA+0v5r/jR5z/8APtL+a/40AS0VF5z/APPtL+a/40ec/wDz7S/mv+NAEtFRec//AD7S/mv+NHnP/wA+0v5r/jQBLRUXnP8A8+0v5r/jR5z/APPtL+a/40AS0VF5z/8APtL+a/40ec//AD7S/mv+NAEteH6//wAjJqf/AF9y/wDoZr2vzn/59pfzX/GvOdT8B6ze6veXUQtwk87yKGk5AZiRnj3rzsdTnUilBXObERlJKxxVFdX/AMK61z/p2/7+n/Cj/hXWuf8ATt/39P8AhXlfVa/8pyexqdjlKK6v/hXWuf8ATt/39P8AhR/wrrXP+nb/AL+n/Cj6rX/lD2NTsei6B/yLemf9ekX/AKAK0KoaYk1lpFnay27l4IEjYqy4JVQDjn2q15z/APPtL+a/419JBWikz047IloqLzn/AOfaX81/xo85/wDn2l/Nf8aoZLRUXnP/AM+0v5r/AI0ec/8Az7S/mv8AjQBLRUXnP/z7S/mv+NHnP/z7S/mv+NAEtFRec/8Az7S/mv8AjR5z/wDPtL+a/wCNAEtFRec//PtL+a/40ec//PtL+a/40AS0VF5z/wDPtL+a/wCNHnP/AM+0v5r/AI0AS0VF5z/8+0v5r/jR5z/8+0v5r/jQBLXl/wAS/wDkZLf/AK9F/wDQ3r0vzn/59pfzX/GuN8X+FtT1/V4rqzSNESARkSuAchmPbPrXHjISnStFXZjXi5Qsjzaiur/4V1rn/Tt/39P+FH/Cutc/6dv+/p/wrxfqtf8AlOH2NTscpRXV/wDCutc/6dv+/p/wo/4V1rn/AE7f9/T/AIUfVa/8oexqdjpfhp/yLdx/19t/6AldhXOeENJvNA0iW1vIt7vOZAYmBGCqjuR6Vvec/wDz7S/mv+Ne/h4uNKKe56FNNQSZLRUXnP8A8+0v5r/jR5z/APPtL+a/41uaEtFRec//AD7S/mv+NHnP/wA+0v5r/jQBLRUXnP8A8+0v5r/jR5z/APPtL+a/40AS0VF5z/8APtL+a/40ec//AD7S/mv+NAEtFRec/wDz7S/mv+NHnP8A8+0v5r/jQBLRUXnP/wA+0v5r/jR5z/8APtL+a/40AS0VF5z/APPtL+a/40ec/wDz7S/mv+NAEtc/45/5Eu+/7Z/+jFrb85/+faX81/xrL8SWVzrHh+5sbaEpLLt2mRlCjDg84J9KyrJypyS7MiavFpHjFFdX/wAK61z/AKdv+/p/wo/4V1rn/Tt/39P+FfPfVa/8p53sanY5Siur/wCFda5/07f9/T/hR/wrrXP+nb/v6f8ACj6rX/lD2NTsS/DT/kZLj/r0b/0NK9QriPCHhbU9A1eW6vEjdHgMYETgnJZT3x6V2XnP/wA+0v5r/jXtYOEoUrSVmd1CLjCzJaKi85/+faX81/xo85/+faX81/xrsNiWiovOf/n2l/Nf8aPOf/n2l/Nf8aAJaKRSSoJUqfQ9qWgAooooAgvbO31GyltL2FZreZSkkbjhgag07R7LS3me0SQyTlfNlmneZ3xwAWck4GTgZwMn1q9RQAUUUUAFFFFABRRRQAUUUUAFFYHifxSPDUlj5lm1xFcuwldXwYUUZZ8Y+bAycccCpx4ktUvNUS6aOC106KKVrkyZV1kUnpj2465zQBsUVjjxVo/2a6nluXgWzCm4W4t5InjDHCko6hsH1xjr6VHN4x0WAJ501yrPEZljNlPvMYzl9uzO3jrjH50AblFZVn4n0e/mkjtb1W8uEzl2RkRowcF1dgFZQeCQSBWfqfjC2TRL650li91arG5iureSI7XcKG2sFJB55HHFAHS0Vl+INTl0vTUa0RHurieO2gVwSu92AycEcAZPXtWa/i6RGe7Onr/ZEd99ie78/wDeBt3l7/L24KeYdud2e+KAOmorM0vU5bvUdUsrpUEtjOFUxggNG6hkPPfkg+4rToAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiuf0PxUmr6nqllNbfZHsJXVWaXcJY1kZC/QY5Q5HOMjmqVn48jutPvro6dKpiuI4bSJZAWuvNAMRGQAu4EHqcA9aAOtorBvdc1LStPmudT0qFdkkKIbe88xG8yVYzklFII3Z+6QfWrWuapLpy2UVqqNcXt2lvHvBIUHLO2AQThVY9euKANSisifW5oPFFppT6eyw3KSFLppV+YooY4UZOOcZOOc8Ec1Jompy6gL6G6Ea3Fldvbv5YIBHDI2CT1Rl/HNAGnRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFMmmjt4JJpmCRxqXdj0AAyTWBpPiz+09Av9QksWtprJDI1s0mSUMYkQ5xxuUjtwc9cUAdFRXO2niS7B0x9W02G1ttU2rbzW90Ztrsu5VcFFIyM4I3DIwcZGb82py/8JRbaVbqm02z3Nw7A5CghUC89SSTz2U0AadFc9/buq/2gmn3Olw2dzeW00tm32nzgGTHyyKAuPvL91mHbPetPRNSXWNDs9QQbftEKuy/3WxyPwOR+FAF6iiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAydU0mS/wBb0q6/dNb2hm85JOS4eMrgDGD15zXPW3ge/trHWbRb2F0me3OntJuYosD70ST2HC8Z+UetdvRQByOp+G9U1qPUbm8+x213cW8VtDDFM0kaoknmFmcoCSSSAAvGOpzxp32i3Fzr1xexvEIpdMazUMTuDlic9On+cVt0UAcsfDeppa6MtrdwwXGn6TNZmUEnErJGFZRjkAxk847cVlHwVrE8d+Z5IEe5s4bdfMv57k7kk3li8i5APPAGB+Oa76igDnvGQ8vT7C+b/VWGowXEvsm7ax/ANn8KpnwxqTQy6QXtRpMuom8aYSN52wyeb5Wzbj7/ABu3dO2a62igDn9CHn+KPEV6nMLTQ2yt2Zo0+bH0LkfUGugoooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAOIvPBOoTwn7New200t9dGd0LfPaXEhZkzj7wG0jtkdau3XhO4km1OW0mht2e5trmx4LKjQxqoVx6HaRwehz7V1VFAHO6pp+t63pM1rdRafaky27xrHO8udkyuxLFFxwuANp570eKV8m80LUm/wBVZ6golJxhVkRo9xPbDOv510VFAHPanY6zceKNPvrS2sWtrESKPNu3V5BIqgnaIiBgg9zn2pfDANxqGvakP9XdX5SIjoyxRpFuH/AlYfhXQUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBleJNMutZ0ObTrSZYBdFY5pCTlYifn28HJK5GDxzWRJ4T1CHUbya11NrqK+057SdbsIhVgP3RURooIG5gcjIB79K6yigDmLbRNWuotHttWWzt7bSmjl/0ad5XndE2rncihF5JP3ieBx3mlX7N8RoJpOEvdNaGM+rxybiv1KuT/AMBNdDRQBy2m2+r6ZeXmseI4bBmMbGa6hu3byIVywSOMxDA9ctknnsALvgy1ltPBumx3CFJWi8xkPVd5LY/DdW5RQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB//2Q==)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z8gNrUKxMzgd"
      },
      "source": [
        "data_1 = data.drop([\"ADAS11\",\"ADAS13\",\"MMSE\"],axis=1)\n",
        "data_copy = data_1.copy()\n",
        "#index 12, 48 featue is outlier"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "daiQkufJD9Vb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 352
        },
        "outputId": "44b84adb-988f-478b-cec9-73776f807ef6"
      },
      "source": [
        "# Normalized Data\n",
        "df = data.drop([\"DX_bl\",\"ADAS11\",\"ADAS13\",\"MMSE\"],axis=1)\n",
        "data_norm = (df - df.mean())/df.std()\n",
        "\n",
        "# Figure Box-plot\n",
        "plt.figure(figsize=(15,5))\n",
        "plt.subplot(121)\n",
        "sns.boxplot(data[\"DX_bl\"],data[\"ST85TA\"],showfliers=True)\n",
        "plt.subplot(122)\n",
        "plt.scatter(data[\"DX_bl\"],data[\"ST85TA\"])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f4d35417da0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZD6FNqa2NpWj"
      },
      "source": [
        "The above figure is a boxplot of one feature (ex: ST85TA). The middle line in the box represents the median, and the box represents the distribution from 25%(Q1) to 75%(Q3). The points outside the boxplot are outliers, and these are far from the distribution of the data. In fact, this data analysis reveals a lot of information while visualizing it, but it should not be assumed that it is an outlier because the brain volumes of subjects are all different.\n",
        "\n",
        "If the distribution is good, the difference between the median values is large, and the boxes should not overlap by class, but these data overlap considerably and the median value also shows little difference. It is indicated that they are not suitable for learning. As I will mention repeatedly, the distribution of these data is a big problem.The outliers thought it could be a hindrance to learning, so through the code below, I dropped data that was 1.5 times the length of the box, 1.5 IQR.          \n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RKfONWBDL9Y4"
      },
      "source": [
        "# Remove outlier function\n",
        "def remove_outlier(data, column, gr):\n",
        "  \n",
        "  Class = data[data[\"DX_bl\"] == gr][column]\n",
        "  q25 = np.percentile(Class, 25)\n",
        "  q75 = np.percentile(Class, 75)\n",
        "\n",
        "  iqr = (q75 - q25) * 1.5\n",
        "  \n",
        "  lowest = q25 - iqr\n",
        "  highest = q75 + iqr\n",
        "  outlier_index = Class[(Class < lowest) | (Class > highest) ].index.values\n",
        "\n",
        "  return outlier_index.tolist()\n",
        "  "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hpNk9zg7Q6Af",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "151c446e-d691-4478-8a95-2d6a2a929142"
      },
      "source": [
        "# Drop the outlier data\n",
        "col_list = list(data_1.columns.values)[1:]\n",
        "out0 = []\n",
        "out1 = []\n",
        "out2 = []\n",
        "\n",
        "for j in range(3):\n",
        "  temp = []\n",
        "  print(\"//////class{}///////\".format(j))\n",
        "  for i in range(len(col_list)):\n",
        "    out = remove_outlier(data_copy, col_list[i], j)\n",
        "    temp.extend(out)\n",
        "\n",
        "  cnt = Counter(temp)\n",
        "  common = cnt.most_common(17)\n",
        "  com_index = [word for word, word_count in common]\n",
        "  \n",
        "  if j == 0:\n",
        "    out0.extend(com_index)\n",
        "  elif j == 1:\n",
        "    out1.extend(com_index)\n",
        "  else :\n",
        "    out2.extend(com_index)\n",
        "\n",
        "  print(common)\n",
        "  print(com_index)\n",
        "#drop the outlier data that we wwant\n",
        "\n",
        "#data_drop = data_copy.drop(out0, axis= 0)\n",
        "data_drop = data_copy.drop(out1, axis= 0)\n",
        "#data_drop = data_copy.drop(out2, axis= 0)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "//////class0///////\n",
            "[(1521, 36), (533, 26), (448, 24), (506, 20), (1304, 19), (578, 15), (845, 15), (1477, 14), (1190, 11), (1383, 10), (143, 10), (1363, 9), (303, 9), (1310, 9), (1443, 8), (1217, 8), (577, 8)]\n",
            "[1521, 533, 448, 506, 1304, 578, 845, 1477, 1190, 1383, 143, 1363, 303, 1310, 1443, 1217, 577]\n",
            "//////class1///////\n",
            "[(1172, 38), (1031, 36), (335, 26), (737, 24), (907, 24), (1378, 23), (990, 22), (895, 22), (461, 20), (1048, 18), (987, 18), (1084, 18), (513, 17), (1472, 17), (1568, 16), (200, 15), (261, 15)]\n",
            "[1172, 1031, 335, 737, 907, 1378, 990, 895, 461, 1048, 987, 1084, 513, 1472, 1568, 200, 261]\n",
            "//////class2///////\n",
            "[(528, 28), (1577, 22), (666, 19), (856, 18), (141, 14), (943, 12), (493, 11), (420, 11), (660, 10), (1226, 10), (332, 9), (1417, 9), (304, 9), (1139, 9), (898, 9), (1423, 8), (1462, 8)]\n",
            "[528, 1577, 666, 856, 141, 943, 493, 420, 660, 1226, 332, 1417, 304, 1139, 898, 1423, 1462]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EuG4vwKE0VU3"
      },
      "source": [
        "# Correlation application\n",
        "corr = data_1.corr(method = 'pearson')\n",
        "aa = corr[corr['DX_bl']< -0.3].T\n",
        "#aaa = correlation이 -0.3보다 큰 feature들의 label\n",
        "aaa = aa.keys()\n",
        "data_corr = data_copy[aaa]\n",
        "data_corr2 = data_drop[aaa]\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fYLd7iVvC3DE"
      },
      "source": [
        "def DecideXY(data):\n",
        "  if data == 'drop_all':\n",
        "    x = data_corr2.iloc[:,1:]\n",
        "    y = data_drop['DX_bl']\n",
        "  #only drop outlier data \n",
        "  elif data == 'drop_row':\n",
        "    x = data_drop.iloc[:,1:]\n",
        "    y = data_drop['DX_bl'] \n",
        "  #only drop uncorrelative feature\n",
        "  elif data == 'drop_col':\n",
        "    x = data_corr.iloc[:,1:]\n",
        "    y = data_1['DX_bl'] \n",
        "  elif data == 'None':\n",
        "    x = data_1.iloc[:,1:]\n",
        "    y = data_1['DX_bl']\n",
        "  \n",
        "  return x, y"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9YmNZoEoGfci"
      },
      "source": [
        "x, y = DecideXY('None')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nHXuMO151ceQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "acad75e0-4d1b-459d-fdec-e76bb32aceb9"
      },
      "source": [
        "# 10-fold Cross validation set\n",
        "skf = StratifiedKFold(10,shuffle=True, random_state = 1)\n",
        "train_index = list(range(0,len(aaa) * 10))\n",
        "valid_index  = list(range(len(aaa) * 10,len(x)))\n",
        "x_train_list = []\n",
        "x_valid_list = []\n",
        "y_train_list = []\n",
        "y_valid_list = []\n",
        "for train_index, valid_index in skf.split(x, y):\n",
        "  x_train, x_valid = x.iloc[train_index, :].values, x.iloc[valid_index, :].values\n",
        "  y_train, y_valid= y.iloc[train_index].values, y.iloc[valid_index].values\n",
        "  x_train_list.append(x_train)\n",
        "  x_valid_list.append(x_valid[:160,:])\n",
        "  y_train_list.append(y_train)\n",
        "  y_valid_list.append(y_valid[:160])\n",
        "  \n",
        "\n",
        "for x_valid in x_valid_list:\n",
        "  print(x_valid.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n",
            "(160, 140)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JDwnSj36cUe8"
      },
      "source": [
        "**1.Drop the outlier data**    \n",
        "We extracted 1.5*IQR data for each feature and listed the index and frequency of the data. The most extracted 17 data were designated as outlier and dropped. (The number of data dropped here is randomly designated during the experiment, so its meaning is not significant.)                \n",
        "You can drop outliers of data of several classes, but there is a problem here. Data that is outlier in one class may have important information in another class. Therefore, it was not possible to drop all the outliers randomly, so only the data that appeared as outliers from many features were arbitrarily removed. And because the data of the MCI group is the largest and it is located between the normal and the AD group, so the outliers of the corresponding group were removed.\n",
        "\n",
        "**2. Correlation between labels and features**    \n",
        "The Pearson correlation coefficient is a quantification of the correlation between two variables. It has a value between +1 and -1, where +1 means a perfect positive linear correlation, 0 means no linear correlation, and -1 means a perfect negative linear correlation.             Although the size of the Pearson correlation coefficient is not absolute, it is usually considered to have a strong correlation if the size exceeds **0.3**.   \n",
        "Features with positive correlation with label were not extracted, and there were **23 features** with strong negative correlation, and we tried to learn through them.\n",
        "\n",
        "If we train with only 23 features, we can drop features that may be important for learning, so we tried learning with 38 features with a pearson correlation coefficient less than **0.25**.\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6hwdQcFP6JB-"
      },
      "source": [
        "# Resampling Function \n",
        "def Resampling(x_train, y_train, sample):\n",
        "  if sample == 'RandomUnderSampler':\n",
        "    sampling = RandomUnderSampler(sampling_strategy='majority')\n",
        "  elif sample == 'TomekLinks':\n",
        "    sampling = TomekLinks(sampling_strategy='majority')\n",
        "  elif sample == 'SMOTETomek':\n",
        "    sampling = SMOTETomek()\n",
        "  elif sample == 'RandomOverSampler':\n",
        "    sampling = RandomOverSampler(sampling_strategy= 'minority')\n",
        "  elif sample == 'SMOTE':\n",
        "    sampling = SMOTE(sampling_strategy= 'minority')\n",
        "\n",
        "  x_train_re, y_train_re = sampling.fit_sample(x_train, y_train)\n",
        "\n",
        "  print(x_train_re.shape, y_train_re.shape)\n",
        "  print(Counter(y_train_re))\n",
        "\n",
        "  return x_train_re, y_train_re\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MQzAf7B-N9p2"
      },
      "source": [
        "**Resampling (Oversampling & Undersampling)**        \n",
        "As seen in the boxplot graph above, the distribution of data according to labels tended to overlap too much. In addition, the MCI group has about twice as much data as other groups, so the data is imbalanced, which makes learning difficult. So, the data was augmented or corrected using various resampling techniques.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YJcCfcLK5O9z"
      },
      "source": [
        "![1.JPG](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAeAB4AAD/4RD6RXhpZgAATU0AKgAAAAgABAE7AAIAAAAQAAAISodpAAQAAAABAAAIWpydAAEAAAAgAAAQ0uocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEd5ZW9uZ055ZW9uIEtpbQAABZADAAIAAAAUAAAQqJAEAAIAAAAUAAAQvJKRAAIAAAADNDQAAJKSAAIAAAADNDQAAOocAAcAAAgMAAAInAAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADIwMjA6MTA6MzEgMTM6MTc6MjAAMjAyMDoxMDozMSAxMzoxNzoyMAAAAEcAeQBlAG8AbgBnAE4AeQBlAG8AbgAgAEsAaQBtAAAA/+ELImh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMjAtMTAtMzFUMTM6MTc6MjAuNDQ0PC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPkd5ZW9uZ055ZW9uIEtpbTwvcmRmOmxpPjwvcmRmOlNlcT4NCgkJCTwvZGM6Y3JlYXRvcj48L3JkZjpEZXNjcmlwdGlvbj48L3JkZjpSREY+PC94OnhtcG1ldGE+DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgIDw/eHBhY2tldCBlbmQ9J3cnPz7/2wBDAAcFBQYFBAcGBQYIBwcIChELCgkJChUPEAwRGBUaGRgVGBcbHichGx0lHRcYIi4iJSgpKywrGiAvMy8qMicqKyr/2wBDAQcICAoJChQLCxQqHBgcKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKir/wAARCADpAr8DASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwD6RooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiomuoEuUt3mjWZwSkZcBmA6kDqaAJaKijuYJZpIopo3kiwJEVgSmemR2qWgAoqvc39paTQQ3VxHFJcNshV2wZGxnA9asUAFFFFABRRRQAUUVXvb+1063NxfTpBECAXkbAyelAFiiq8GoWlzdzWsFxHJPAFaWNWyyBhlSR2yKsUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUyaaK3haWeRIo1GWd2AAHuTVey1Sw1JWOn3tvchephlD4/I0AW6KKKACiiigAooooAKKKr31/a6bbG4v7iO3hBCmSRsDJOAM0AWKKrwX9rc3M9vbzxyTW5AmRWyYyRkAjtxVigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACvN/EB2/tBeFCxwDpt0Bnua9Irm/FvgbSvGP2SS/e6tbuycvbXllMYpYiRggMOxoA80vNbvfD2s/FjV9KkCXVmts8TsoYK3lL2PFXl8a+KvDfiJE12/h1S2u9Am1RYY7cR+TJGu7aCOSCPWuws/hhoFp4V1TQt15NFq3N7czT755j6lyO1aE3gnSLjWLPUp0lkms7J7GNGYFGiYYIYY5NAHkr3PiTUda+HGseIdWt71NRvWuI4YrcR/ZyYidoI+8MevOan0f4neKdU1GHVrX7ZdWk2pG3OlRaPK0aW4cqX+0BcbxjJ5x29q7nS/hF4e0rUNPu4bjUpTpspks4prsvHBkEbVUjpzU0Xwr0GDWhfRTagkAuzeDTxckWwmPVtn9M4oA5XT/EXjzV7fxRqWm3sEy6RcT29rp4tl3XDD7pLcdPTvUPg34i31zJq0Or+IjHc29gZza6zpptJbdx95vl4aMenWvQrfwLo1vpOraaEme21eV5rlXk53N12kYxWXb/CjQlF1/aFzqOqNcWbWQe+ud5ihPVV4GPr1oA4rw74/8Rt4kutPuNVm1K3k0OfUIbmfSzaASJjBjDAF056kVDpXi3x5qH/CIStrlqg8TWzhk+xri2Kj74/vE9cHjmu8034V6Lp14t39t1O6nWyewD3FzvxC4xtHHGMcVes/h/otiugLB9oxoCMlnukB4Iwd3HP6UAecr8QvFreFtFSO/tv7TuPEh0eW5eAbJEyyhivbseMdKz/Geu65/ZPiLwv4jvo9TksJ7O4hvI4REWR5B8rKOMg11ni74XRz2eh6dokU8loPEC6hfFptrIhzvKsMEYzxjmtyH4U6AmlX1ncTX909/NHNcXVxcb5nKHKjcR0HpQBxeteK9S0PVPiFc6WLeG402w054JRApYl153HHzD0z0q9H4y8T+H/EskOtX0Op29xocmqRRLbiPyXRc7ARyRzjJrsdQ+HWh6k+ttc/ac63DBDd7ZAPli+5t44PrU974M0+W/TVI42lvbfT3sYUlf8AdujDGGGOegoA8707xb40gTwRqOoa3a3Nt4nvY/MtltVUwRspYIp7jBGT1yPevaK8K8OfDjWP+El8PmbRLjTYdJvjdSyz35niRQDiOBf4QSQT9B6V7rQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeXfF3/iYeJfBPh++kZNH1TUJPtqhiolKKCkZPoxJGKy/Eug2fg/4meGI/AsUWlXGprNbzw264jKhCVdk6HBxzXpnijwppXjDSf7P1qBpI1cSRyRttkicdGVh0NZXhz4baP4d1g6t9o1DU9R2eWlzqVyZ3jX0XPSgDjdD+JOua7e+G9GhdE1ONrmXXv3YykduduMfw72x06Vn+FPiV4p1nVNM1H/AEy8sdRumilsU0eVY7WLJCuJ9uGIwM84r0/SvA+iaN4p1bxBY27Lf6sALlmOV45O0dsnk+pArM074W6FpetRX9tNfmKCZp4LFrkm3idupVP6dKAON0fxN8QdX8Iax4isr2K5azluLW30+O0UtKVk/wBZnuQpICjrtHrU3g/4hXl5puvfb/EifabK085Y9X082ktqccs4Xhlz2HPbvXdW3gDRLbwreeHkFw1ldzPPJmX5w7NuJDDGMHkVQt/hVoK2+ox6hNf6m+oWwtZZr243usY5CqcDHPNAHn+n/ETxRbXmo202rS6hGdFkvoLi40v7JiRT1RWALRnsSKuaf4j8d6hr+g6e2v20a6/pjXW4WSn7IV/u8/Nn39a7C3+E2hxSGW4vdUu5ms2sWluLncTC38PTjHb9c1r2ngfSbPU9JvofP87SbVrS2y4I2N13ccmgDzOz+IXi3U/DPhFYL+2g1HVNRmsLm4aAFW2FgG29jx0GKzPGmv6zeeBPFGgeIrqO/utH1SzRbyOIR+ajyqRlRwCMGus8RfCqNj4Y0vRUuf7Ns9RkubmXz9skQcEllYY5ya6BPhV4fHhq70aV72dL25S6ubmWfdNLIpBBLEew4oA4jWfGGq6JqPjqfSRbQ3Fnf2MUb+QuXDoud5x83XGT0q8/jfxN4W17xHZa1ew6ulpoJ1i2ItxFsYHb5fHUZ712V/8ADjQtRk1Z7j7TnVp4Z7nbIB80QAXHHA4qTVfBdjPfahrFvB9o1G40ptNEM7/uZI85AYYzye/pQBxGj+JfGNprvg2PV9btr628QoZ5Y0tVjaMGPcEBHYZHPWvX68R8GfDzV4fF2h3l1o9zpsGk+Y0j3V/9oBJGBHEP4UGc17dQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFZXie+n03wzf3lowSaGIsjEZwfpWrWL4whluPCGpRW8byyPAQqIpJY+wFXTtzq/czqXUHbseM/8LU8V/8AP7H/AN+E/wAKP+FqeK/+f2P/AL8J/hWF/wAIzrv/AEBdQ/8AAV/8KP8AhGdd/wCgLqH/AICv/hX0/ssN2X4Hy3tsV3f4m7/wtTxX/wA/sf8A34T/AAo/4Wp4r/5/Y/8Avwn+FYX/AAjOu/8AQF1D/wABX/wo/wCEZ13/AKAuof8AgK/+FHssN2X4B7bFd3+Ju/8AC1PFf/P7H/34T/CvUvhxr1/4i8NyXeqyrLMs7ICqBeAB2H1rxD/hGdd/6Auof+Ar/wCFeyfCexu7DwnLFfW01tIbliEmjKHGBzg1w46FGNG8Er+R34CpWlWtNu1up3NFFFeGe8FFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFcP468fXHhHULa3gs47gTRlyXYjHOK5X/AIXXff8AQKg/77NdcMHWqRUorRnHUxtCnJwk9UexUV47/wALrvv+gVB/32aP+F133/QKg/77NX9QxHYj+0MP3/A9iorx3/hdd9/0CoP++zVnT/jFeXmpW9s2mQKJpFQkOeMnFDwFdK9gWYYdu1/wPWaKKK4TvCiiigAooooAKKKxB4v0Q6kLIXbeYZfJEnkv5Rk/ueZt2bvbOaANuiiorm6hs4DNcuEjBALEdyQB+pFAEtFFFABRQSAMk4FULPXNNvrGa9trtDawbvMnYFEUL1OTgYHr0oAv0VXt7+2urieG3lDyW7BZVAPykjI/Q1YoAKKKKACimu22Nm9BmuCf4i3Kuy/YouDj7xrjxWNoYW3tXa52YbB1sVf2SvY7+ivPv+Fj3P8Az4xf99Gj/hY9z/z4xf8AfRri/trBfzfgzr/sbGfy/ij0GivPv+Fj3P8Az4xf99Gj/hY9z/z4xf8AfRo/trBfzfgw/sbGfy/ij0Giua8MeKZdeu5oZbdIhGgYFSTnmulr0sPiKeIp+0pu6PPr0KmHn7OotQoryK4+M17DdSxDS4CEcrneexqL/hdd9/0CoP8Avs16v1DEdjyf7Qw/f8D2KivHf+F133/QKg/77NH/AAuu+/6BUH/fZo+oYjsH9oYfv+B7FRXjv/C677/oFQf99muu8B+OZ/F9xexz2kduLZUIKEnOSf8ACoqYOtTi5SWhdPG0aklCL1Z2lFFFch2BRRRQAUUUUAFFFFABRRRQAVjeL5BF4Q1J2jWULASUcnDexwQfyNbNYXjb/kStU/64NWlL+JH1M6v8OXozwD+2bX/oB6f/AN9z/wDx2j+2bX/oB6f/AN9z/wDx2sqivrORf02fHe0f9JGr/bNr/wBAPT/++5//AI7R/bNr/wBAPT/++5//AI7WVRRyL+mw9o/6SNX+2bX/AKAen/8Afc//AMdr034eXclv4ffV7W2SG1E5ju7aFnYBQBiRQzMcjPI7j6V47XuPwdAbwZMCMg3T5H/AVrgzCKVD5no5bJuvbyO+jkSWNZImDowyrA5BFOrEtSdD1EWMh/0G4Y/ZmPSNupj+h6j8q26+ePpAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDxn40/wDIe0//AK9z/wChV5pXpfxp/wCQ9p//AF7n/wBCrzSvqcH/AAInyeN/3iQUUUV1HGFX9C/5GCw/6+E/9CFUKv6F/wAjBYf9fCf+hCpn8LKh8SPqIdKKB0or44+2CiiigAooooAbKpaF1XqVIFcDoGsWWneF9N8P3lg91qluyQS2jQ5JcNzLkjGP48+9egUnGc8ZoA8xhvL1Vs5IrzUTqzfaP7XieWQxxIIpDkITtTEgiClQOCfU1UtL68fSYJND1DU7pXtLR7x5p5XIlM0ecbj8pKlsheMV6gHtdUtJ41bfEWeCTAI5BKsPzzUlrbRWVnDbW67YoUWNBnoAMD9BQB5hBda2+uTeZezw332+dZIxPO+IAW2gQ48sDZtIYc5wc5JrZ+H11dNfXNvcTXF0BbxubhrqWWNmyQfll5R+5UEiu0u722sRG104QSyLEhwTlicAcU+2uILuBZrWRJYmzh0OQecdaAPPJ5buXUJUe61D+2W1WSJ7dZpViFnvIUhAdgHlbW3DndnnNTjSCvwJ1CxiS5Mr6ZclUaWQuX2sQM5zjOOOh6Ywa9B461XS+tpL+SySUG4iRZHTB4Vs4P6GgDzdbM7tV1XSrnUQY7m0NqUuZdroUQEsuf3mR13ZpdHutZk8X2/2i7mS7OoSrcQNcTsDDtbjyseWq42kMOc455Nen8UcUAFFFQ3V1HaW7TTE4HAAGSxPQAdyTQBFqN2lpatkF5HBVI16sa8Xk/1rZ45NexwWsjCS7vcCd0IVc5ES+n+NeVS2EHmt/wATO16n+GX/AOIr5XiGDl7O3n19D6fIJKPtL+X6mdRV77Bb/wDQUtP++Zf/AIij7Bb/APQUtP8AvmX/AOIr5X2MvL71/mfUe1j5/c/8ijRV77Bb/wDQUtP++Zf/AIij7Bb/APQUtP8AvmX/AOIo9jLy+9f5h7WPn9z/AMjpPhz/AMhS7/65D+deiHpXB+AbaOHUbkx3cM5MYyIw4xz/ALSiu8PSvvMli44OKfd/mfD5xJSxcmvL8j5Wv/8AkI3H/XVv51Xqxf8A/IRuP+urfzqvX6Utj8xe4UUUUxBXqXwS/wCP7V/+ucX82ry2vUvgl/x/av8A9c4v5tXHjv8Ad5f11O3Af7zH+uh6/RRRXzB9WFFFFABRRRQAUUUUAFFFFABWN4vi87whqUe9E3QEbnOAPqa2axfGE0lv4Q1KWFykiQEqw6g1pT+NepnV/hyv2Pn3+wj/ANBPTv8AwI/+tR/YR/6Cenf+BH/1qT/hJdZ/6CM//fVH/CS6z/0EZ/8AvqvqrVf6/wCGPkr0vP8Ar5i/2Ef+gnp3/gR/9aj+wj/0E9O/8CP/AK1J/wAJLrP/AEEZ/wDvqj/hJdZ/6CM//fVFqv8AX/DBel5/18xf7CP/AEE9O/8AAj/61ex/Ce3W08LTQi4gnYXDMTC+4DIH+FeN/wDCS6z/ANBGf/vqvWPho97qPhZ777Qz3sVwyq0h4dcA7G9vftXBj+f2Pvdz0Mv5Pb+72O9vrKLULN7ecZVhwR1U9iPcGqekXspeTTtQP+m24GW6CZOzj+voau2l2l3DvUMjKdrxt95G9DVXV9Pe6jjuLMiO9tiWhc9D6qfY9K8A+hNGiqmm6gmpWYmRSjg7ZI26xsOqmrdABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeM/Gn/kPaf/17n/0KvNK9L+NP/Ie0/wD69z/6FXmlfU4P+BE+Txv+8SCiiiuo4wq/oX/IwWH/AF8J/wChCqFX9C/5GCw/6+E/9CFTP4WVD4kfUQ6UUDpRXxx9sFFFFABRRRQBjeMGuU8E621g0q3K2E5haHO8P5ZxtxznPTFcdf6TeWV3qb2t1qzfZVs5LcG7mYF2kxIcZ546joPQV6VRQB5xBaalqmvwW9/PqS2gOruwjnkiDFbmEQglSDgKW2jPTPbNZ0VzqbafK2oXGqDVFtYBpqq8gV26HcB8rHON27tXrFGKAPODZ6kyz3kk+pG4fWxb7PPkCCDzAcBQcY68+hxnFZ9ml/pmh2Ntpc93GdUWTTmWSWQmCbzCfMXceCELnI4+UV6vVObS7WfU4L+ZC08ClYyWO1c9Tt6Z9+tAGL4qEkFvpUbSXKacLgLevA7h9gQ7csvzY3bckc1yU80tu97NapqbW8sVrEk0sjq23zn5ZwC+3HUj5sH8a9T69aMD0oA8ijm1R9Nuklu7qKCLVHEIaW58t4zAhx5g/eAby2CeM1NHe6l/aGnXLf2k0jC1/wBDknmWRASoO0geXIDzu3gHr7V6vgelIxVFLPgBRkk9qAGyzJbwtLM21FGSTVK1hkvLhb27UqF/1EJ/gB/iP+0f0HHc02FW1SZbmUEWiHMKH/lof759vT86uXN1FaQGWdsKOAByWPYAdz7UAJeTx21nJJM4RQp5NeJyHMrEepr2FLWW6zdX64bafKg6iP6+rV5XLo2pGVsWFz1P/LI18rxDCUvZ8qvv+h9PkE4x9pzO236mfRV7+xdT/wCfC5/79NR/Yup/8+Fz/wB+mr5X2NX+V/cfUe2p/wAy+8o0Ve/sXU/+fC5/79NR/Yup/wDPhc/9+mo9jV/lf3B7an/MvvOk+HP/ACFLv/rkP516IelcH4BsLu01G6a6tpYVaIAF0Izz713h6V95ksXHBxTXV/mfD5xJSxcmn2/I+Vr/AP5CNx/11b+dV6sX/wDyEbj/AK6t/Oq9fpS2PzF7hRRRTEFepfBL/j+1f/rnF/Nq8tr1L4Jf8f2r/wDXOL+bVx47/d5f11O3Af7zH+uh6/RRRXzB9WFFFFABRRRQAUUUUAFFFFABWF42/wCRK1T/AK4NW7WF42/5ErVP+uDVpS/iR9UZVv4cvRnzXRRRX158YFFFFABXuPwc/wCROm/6+m/kteHV7j8HP+ROm/6+m/ktefmP8D5npZb/AB/kdld20kc32yyA84DDp2lX0+voasWtzHdwCWI5HQg9VPcH3qaqF1BJa3BvbJSxP+vhB4lHqP8AaH6jg9iPnD6YqajG+k3p1W2UmFhi8jUdV/vgeo7+1bEciSxrJGwZGAZWByCDTIZoru3WSIh43H+QayLUnQdRFk5/4l9yxNsx/wCWLnkxn2PVfxHpQBuUUUUAFFFcn4yS/ku7LyEvrmxUO1za6bdeRcN/dYEMpKjuAfzoA6yiuEtfF/8AZvh+y+zXCX295kaTVrg2skRRseWw2szOM4zjsSTyMh+Il7LavdWWhxywww28sxe82MDKxXao2EHB7kjj8qAO7orkP+Ex1Pzlsv7GtzqJ1P7AYxenylzbfaA+/wAvONuARt65698258cva6gb67jeFLW3kjntFmyhmEioPmIHGWHzEcA9KAPQaK4Q/EeVbebOn2ryQyRB5o7xmtkRw3zNJ5eQQUII2915weN5/Eoj8IrrTW6SPJsVIYZxIru7hEAccYLMOccA9O1AG7RXA+I/EGtzR/2SltFZX63losxt75trQzOwBWTywQcowPyjHXmu7iUpCitklVAOWz+vegB9FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB4z8af+Q9p//Xuf/Qq80r0v40/8h7T/APr3P/oVeaV9Tg/4ET5PG/7xIKKKK6jjCr+hf8jBYf8AXwn/AKEKoVf0L/kYLD/r4T/0IVM/hZUPiR9RDpRQOlFfHH2wUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAZx1rMP8AxN5tv/LjGef+m59P90fr9KdO7ajcNaQMRAhxcSDv/sD+tT3FzFYQIioWY/LFDGOWPoB/XoKAH3V3HZw7nySeERRksfQCoLa0kknF5f4M2P3cY+7ED6ep96W0s3Ev2q9YPcsMAD7sQ/ur/U9/0qS+v7fTrfzrqTaM4UDksfQDuaAJZ2C28hYgAKck9uK8QlOZXwe5r1V7O71pGm1VWt7MAmOyB5f3kP8A7L+deVSLiRgBgAnFfJ8R/wDLv5/ofVcP/wDLz5fqNzRmjFGK+SPqQzRmjFGKAOx+HP8AyFLv/rkP516Ieled/Dn/AJCl3/1yH869EPSvv8j/ANyj6v8AM+Ezr/fJei/I+Vr/AP5CNx/11b+dV6sX/wDyEbj/AK6t/Oq9fpi2PzB7hRRRTEFepfBL/j+1f/rnF/Nq8tr1L4Jf8f2r/wDXOL+bVx47/d5f11O3Af7zH+uh6/RRRXzB9WFFFFABRRRQAUUUUAFFFFABWF42/wCRK1T/AK4NW7WL4vhkuPCGpRQqXkeAhVHc1pS/iR9UZ1f4cvRnzRRWp/wjWsf8+Ev6Uf8ACNax/wA+Ev6V9b7SHc+O9nPszLorU/4RrWP+fCX9KP8AhGtY/wCfCX9KPaQ7h7OfZmXXuPwc/wCROm/6+m/kteRf8I1rH/PhL+leyfCezuLDwpLDdxGKT7SzbW64wK8/MJRdCyfU9HLYSVe7XQ7miiivnj6Qzp4n0+drq2UtCxzPEoyf94D19u9T3MFvq2nGNiJIZlyrKfyINWqzZEbSpmnhBNnIcyxj/lkT1dfY9x+PrkAbpN7KJH03UD/pcAyG/wCeydmH9fetSs3VLFr6CK6sHVbyA77eTs3qp/2SOP17VPpmoJqVmJVUxyKSksTfejcdVNAFusnWPD1trE8Fw01xa3UAIjuLWTY4B6jPccdKTxDrE+j29obS1W5mu7pLZFeTYoLA4JODwMc8dKxbHxnqElzEt/pkEML3E1qzR3RYiSPPIBUDacdSQfagC8PBGmxi3aCa7hmhWVTOkv7yQSMGfccc5Kg/hT7fwXpFrYy2kKSiGVYkYGQk4jbcnP1rm7zx7JdWV/byQxptsHukksL0lhsdFZSxQYPzryMjrWwPGNyHjuH0+MadNdvZwyif94XXcMlNuApKMOpPtQBLrHhBL7UbW5tJpbcnUhe3LpJtbItWgG30/g4+tT/8IVo5gMUkcsmY2jZmk+ZtzbixP97IBz7Vy2q+MtafRo9QhtUsrifSbu8tVW6LoEQwYLqUwX/eHHUDkd+OiPiyW3vm066s1F950MUcaS5EiuMlwSBwAG7dh60AWP8AhELby23ajqTTGRXE5uTvG0EADtjDHjHOasp4b05PD7aN5bG1bkkt827O7dn1zzn1FQarrl9BrK6ZpFjDdTrbG6lM05iULu2gA7Tkk59Bx1rn7fxdeWVvctN9nZ5NRnjUXlzsESqFIQbVYt1xwMe9AHQw+E7BCXmkuLmdpopmnmky7GI5QZ9ASePc1uVwp+IkzRadKNPhhju4FkLXFwUUuWKlEbaQSMZ5K5yK0NJ8Yyaj4mk0uW1hgALhVac+b8vQlCoBBGfuk0AdVRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeM/Gn/kPaf/17n/0KvNK9L+NP/Ie0/wD69z/6FXmlfU4P+BE+Txv+8SCiiiuo4wq/oX/IwWH/AF8J/wChCqFX9C/5GCw/6+E/9CFTP4WVD4kfUQ6UUDpRXxx9sFFFFABRRRQAUUUUAFFFFABRRRQAVQu55Lic2Vk21sfvpR/yzHoP9o/p1p17dOJFtbPDXMgzk8iNf7x/zzUZdNMhS1tUM91JkhSeWPdmPYe/5UAPklh0y3jtraPdIwxFEvU+pPt6mnWdmY3Nxct5ly4wW7KP7q+gotbRbQPcXMgkuHGZZm4H0A7KPT88nmqL391rDtBox8m3BxJfMufwjB6n3PA96ALF/q4gn+x2Mf2q9YZEanhB6uew/nSWOkeXcfbNRk+1XpHDkfLH7IO3161ZsNOttNg8q2QjJ3O7Hc8jdyzHkmi+1K106MNdShS3CoOWY+gA5NAE8pxC/wDumvHJNXvBKwEidT/yyX/CvSZRqurwuWLaXaFTwMGeQfXog/M/SvKJBiRh1we9fK8QzlH2fK7b/ofT5BCMvacyvt+pc/ti9/56J/36T/Cj+2L3/non/fpP8Ko0V8r7ar/M/vPqPY0/5V9xe/ti9/56J/36T/Cj+2L3/non/fpP8Ko0Ue2q/wAz+8PY0/5V9x3XgG+nutSuVmZSBECMIB39hXeHpXnfw5/5Cl3/ANch/OvRD0r7zJZOWDi2+r/M+HziKji5JLt+R8rX/wDyEbj/AK6t/Oq9WL//AJCNx/11b+dV6/SlsfmL3CiiimIK9S+CX/H9q/8A1zi/m1eW16l8Ev8Aj+1f/rnF/Nq48d/u8v66nbgP95j/AF0PX6KKK+YPqwooooAKKKKACiiigAooooAKxfF8vk+ENSkCI+2Ana4yD9RW1WF42/5ErVP+uDVpS/iR9UZ1f4cvQ8A/t1v+gfp//gMKP7db/oH6f/4DCsqivreSPY+O9pLuav8Abrf9A/T/APwGFH9ut/0D9P8A/AYVlUUckewe0l3NX+3W/wCgfp//AIDCvU/hvanVfDUl7bsthdpOyK9su1WGBwy9DXi9e4/Bz/kTpv8Ar6b+S15+YRSoXXc9HLpydez7HTx6zLYyLDr0S25J2rdJzC59z/AfY/nWyCCAQcg9CKbJGksbJKiujDBVhkEVjnSrrSzv0KUeSOtjKf3f/AD1T6dPYV88fSG1QQCMHkVn2OswXcnkSq1tdD70Eow34eo+laFAGYc6RJk5Ni55/wCmB9f93+X0qDUon0y9/ta1UtGQBdxL/Eo/jA9R/KtllDqVYAqRggjrWamdKkEEvzWLnEbHnyv9k/7Pp6dKAJrmztNXjtJZP3iQyrcwsrcbgDg+/WqcnhXS5YjHJE7IZpZyN55aQEN+HJpls39hagto5/4l90/+jv2ic/8ALM+x7flW5QBxereA4205otKLyXElubISXVwxEMDMrHAxzgouB9ea2F8I6Yt79o/fFfMaZYDKfLSRhywXoDyfzrcooAxJvCOlT2UFrJFJ5UFlJYx4kORFJs3DPr+7Xn2qtB4bkPi6DU7nY0Vjam3tnLlpJMnlnz3H9TXSUUAZWq+HbPV7lLiZ54ZljMJe3lKFkJBKnHUZFU/+EK0pGR7b7RbSRs7K8MxVhvxuGfT5R+VdDRQBzf8AwgukLaxWsRuo7aOIQmFJ22ugYsAw74JNWrXwrp9rqkV9GZ2aBnaGN5SyRFhglQenBI/GtqigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPGfjT/wAh7T/+vc/+hV5pXpfxp/5D2n/9e5/9CrzSvqcH/AifJ43/AHiQUUUV1HGFX9C/5GCw/wCvhP8A0IVQq/oX/IwWH/Xwn/oQqZ/CyofEj6iHSigdKK+OPtgooooAKKKKACiiigAooooAKq3t2bcLHCvmXMvEafzJ9hT7y7W0h3EFnY7UQdXb0FZqGaOdlj2y6jMP3jnlYF7D8PTvQBIM2J+z2/8ApGoT/PI7du24+g7AU9ntNEtWuL2bMkhAZyMvKx6KoHJ9gKrvdRaWxs7FGvNRl+Z+eSf7znsP6cCprLSvKn/tDVJRcXmDhzwkI7hB2Hv1NAEC2d3rTCXVFa2s85SzDfM3vIR/6CK1JprbT7MyTvHb28S8sxCqorOk1p7uQw6JD9qYHDTscRJ+Pf8ACnW+hq1wl1q0pvrlDlN4/dxH/YToD78n3oAj+36hq3y6VEbW2P8Ay93CYZh/sIefxP5VbsdHtrKQzHdPct96eU7nP+H0FX6pXmr2NidtxcKJO0a/Mx/Ac0AWpf8AUv8A7prx2W7s/Nb/AIlsfU/8tX/xr0qXUNUvoX+wWIt4yp/e3ZwceyDn88V5NJnzGz1zzXyvEM3H2dvPovI+nyCKl7S/l19S59ssv+gbH/39f/Gj7ZZf9A2P/v6/+NUaK+V9rLy+5f5H1Hso+f3v/MvfbLL/AKBsf/f1/wDGj7ZZf9A2P/v6/wDjVGij2svL7l/kHso+f3v/ADO68AzwS6jdCC0WAiIZIdmzz713h6V538Of+Qpd/wDXIfzr0Q9K+8yWTlg4t93+Z8PnCUcXJLy/I+Vr/wD5CNx/11b+dV6sX/8AyEbj/rq386r1+lLY/MXuFFFFMQV6l8Ev+P7V/wDrnF/Nq8tr1L4Jf8f2r/8AXOL+bVx47/d5f11O3Af7zH+uh6/RRRXzB9WFFFFABRRRQAUUUUAFFFFABWF42/5ErVP+uDVu1k+KYVuPC2oRPu2vCQdpAP68VpT0mvUzq605eh8yUV1P/CNWn928/wC/sH/xdH/CNWn928/7+wf/ABdfVe2gfJewmctRXU/8I1af3bz/AL+wf/F0f8I1af3bz/v7B/8AF0e2gHsJnLV7j8HP+ROm/wCvpv5LXm//AAjVp/dvP+/sH/xdekeAJotC8PyQfZLxoTOzebhJMHA67Ca4MfUjKjZdz0MvpSjXu+x6BRVK11nT7xtsF3GX/uMdrfkeau14B9CVb7TbXUYtl1EGxyrDhlPqD1FZ2/U9G/1ofU7IfxKP38Y9x/GPpz9a26KAK9nfW2oW4ms5llQ91PQ+hHY+xqaSNJo2jkUMrDBB71nXmixzTm6spGs7z/nrF0f2ZejfzqJNYmsXEOuxCHst1HzE31/u/jQATQIsZ0vUgZLSf5YZSeQf7pPYjsfb1p+k3cySvpmoNuuoBlJD/wAt4+zfXsff61oSxQ3lsY5AssUg9eD+NYl7bTbooWkxdwHdZ3Lfx+qN7kce/WgDoKKp6ZqC6jZiUKUkU7JYz1Rh1Bq5QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHjPxp/5D2n/9e5/9CrzSvS/jT/yHtP8A+vc/+hV5pX1OD/gRPk8b/vEgooorqOMKv6F/yMFh/wBfCf8AoQqhV/Qv+RgsP+vhP/QhUz+FlQ+JH1EOlFA6UV8cfbBRRRQAUUUUAFFFFABUVzcx2lu00xIVewGST2AHc06WVIIWllYKiDLE9hXM3mp+fN9olcwxxnbHxkofYd3P6UAW1eee8ZvlF3j52PKWaddvu56//WxlkEst6ptdCJjtwf32oOMlz32f3j/tdB2zVOQx+THHqe63tW5i06Mlprg+r45OT2/M1ora6hqMYW4b+zLIDAghIEhHuw+7+FACJdWOjsbHS4Xvb5vmeOM7mJP8Ujnhfx/AVIukXOosJNdmDrnItISREv17sfrx7U2HUtJ0xPsmkxee4P8Aq7Vd3Pu3TP1Oaf8A8Tu/7xabEfbzJP8AAUAaLyWun22ZHitoEHViFVRWd/bzXZ26LYzXv/TZv3UI/wCBNyf+Ag1Jb+H7KOUTXAe8nHIluW3kH2HQfgBWoAB0oAx/7N1K+51PUPLQ/wDLCzBQfix+Y/pV2z0qysB/otuiHu2MsfqetWXkSNS0jqijqWOBVGTXtMjOPtaSH0izJ/6DmgC7L/qXz02mvHZIdM81s3l11P8Ay7L/APHK9Ol1lpYXFrp93KNp+Zk2D9a8jk5lbPHJr5XiGSXs7q+/fyPqMgi37Sztt+pb8nS/+fy6/wDAVf8A45R5Ol/8/l1/4Cr/APHKo0V8r7SP8q/H/M+n5JfzP8P8i95Ol/8AP5df+Aq//HKPJ0v/AJ/Lr/wFX/45VGij2kf5V+P+Yckv5n+H+R3XgFLRNRufss80reWMiSEJjn2Y13h6V538Of8AkKXf/XIfzr0Q9K+8yVp4OLStq/zPh84TWLkm77fkfK1//wAhG4/66t/Oq9WL/wD5CNx/11b+dV6/SlsfmL3CiiimIK9S+CX/AB/av/1zi/m1eW16l8Ev+P7V/wDrnF/Nq48d/u8v66nbgP8AeY/10PX6KKK+YPqwooooAztS17TtIkRNQnMTOMqNjNkfgKpf8JroP/P6f+/L/wCFc58R/wDj/tP+uZ/nXFV8njs5xGHxEqUErL1/zPqcFk9DEYeNWTd36f5HrH/Ca6D/AM/p/wC/L/4Uf8JroP8Az+n/AL8v/hXk9Fcf+sGK/lj9z/zOv+wcL/NL71/kesf8JroP/P6f+/L/AOFOTxloTuqLeEsxwB5T/wCFeS1Naf8AH7D/ANdF/nTjxBim0uWP3P8AzFLIcMlfml96/wAj3AHIBHesLxr/AMiVqn/XBq20/wBWv0FY/i9Y38Iaks0hjQwHc4XdgfSvuqXxx9UfEVv4cvRnzRmjNav2HSf+gtJ/4CH/AOKo+w6T/wBBaT/wEP8A8VX13Ov6TPjuR/00ZWaM1q/YdJ/6C0n/AICH/wCKo+w6T/0FpP8AwEP/AMVRzr+kw5H/AE0ZWa9x+Dn/ACJ03/X038lryP7DpP8A0FpP/AQ//FV6v8MtQ0zTPDMkK3cswNwzb/srgDgcZAI/WvPzCSdD5no5dFqv8ju7rTbK9UrdWsUo/wBpAapf2AIP+Qbf3dl6Kkm9P++XyB+GKmXxBpJODqNuh9JJAv8AOrUV9aT48i5hkz02SA/yr54+kM7dr9p1Sz1FB3UmB/yO5T+Ypf8AhIoITjUbW7sD3M0WU/77XK/rWvmgjPWgCC3vbW7QNa3EUqnoUcGpXjSVCkihlYYIIyDVG50LTbly72qJIf8AlpH8jfmMGoP7Jvbbmw1SYDtHcASL+fWgBjaTdaaxl0GVQmctZTE+U3+6eSh/Me1SRajaaqrWV5G9rdYyYJvlYH1U9D9RSfbNZtf+PmwjulH8dtJg/wDfLf41Bd6lo2oRiLVUe1YHKm5jMZQ+ofoD9DQBVuJLjR9S+1sC7hcXKqP+PmIf8tFH99e47j8Kib4n+ElYg6m2Qcf8e8n/AMTU9zHdw2WN51SxHzRzREGaH0YHo2K8hudUi+1S/wDE5UfOeDpacc124WhGs3zdP67M4cXiJUUuXr/XdHq//C0PCX/QTb/wGk/+Jo/4Wh4S/wCgm3/gNJ/8TXkn9qRf9Bpf/BWlH9qRf9Bpf/BWld31Cn5/1/26cH9oVfL+v+3j1v8A4Wh4S/6Cbf8AgNJ/8TR/wtDwl/0E2/8AAaT/AOJryT+1Iv8AoNL/AOCtKP7Ui/6DS/8AgrSj6hT8/wCv+3Q/tCr5f1/28e36J4t0bxHNJFpF2Z3iXc4MTrgdP4gK2a8z+Ft2lzqd6EvxdYhBwLRYcc+o616ZXl4imqVRxX9fkerhqjq01KX9fmY1z4s0azupLe4uyksZ2svlMcH8BUX/AAmug/8AP6f+/L/4V554l/5GW+/66msqvha2fYmFSUEo6NrZ/wCZ9xRyPDTpxk29Uu3+R6x/wmug/wDP6f8Avy/+FH/Ca6D/AM/p/wC/L/4V5PRWX+sGK/lj9z/zNP7Bwv8ANL71/kesf8JroP8Az+n/AL8v/hVmw8S6Vqd0LeyuTJKQSF8th0+orx6ul8Cf8jPH/uN/KujC53ia1eFOSVm0uv8Amc+JyXD0aEqkW7pX6f5HqNFFFfYHyYUUUUAFFFFABRRRQAUUUUAeM/Gn/kPaf/17n/0KvNK9L+NP/Ie0/wD69z/6FXmlfU4P+BE+Txv+8SCiiiuo4wq/oX/IwWH/AF8J/wChCqFX9C/5GCw/6+E/9CFTP4WVD4kfUQ6VWv8AULbTLU3N7J5cQIBbaT1+lWR0rnPHf/Iry/76/wA6+GxVWVGhOpHdJs++wtNVa0Kctm0iX/hNdB/5/T/35f8Awo/4TXQf+f0/9+X/AMK8nor4/wD1gxX8sfuf+Z9Z/YOF/ml96/yPWP8AhNdB/wCf0/8Afl/8KP8AhNdB/wCf0/8Afl/8K8noo/1gxX8sfuf+Yf2Dhf5pfev8j1j/AITXQf8An9P/AH5f/Cta1vre8slu7eTdA4JDkEcD6/SvEa09Z17WJPClno2iWF6U2Hz7iOByG+Y/KCBXrZVmlbGVnColZK+l+68zy8zy2jhKSnTbu3bX/hjp9b8UR3935Nq263iPHzbQ5H8RPYeh/IE42tty0eyaedLTIws8qZkx6Qxckf7zZJ9K5vw/ps1tEjXEVwj9dzQSAg/98HH4V1tk8VsS0cotnb70kenzO5+rMua+iPny/YxXK7jo2mGJpPv32psd7++37x+h21Zk0qB/n17Umu8cmN2EcQ/4AOv45qp5+nyf8feoarP6r5Eqj9EFSxS+HImDCzkdh0aSyldh+JUmgC5FrOj2yeVYsJAvGy0iaTH/AHyDTv7WvJv+PPR7lvRp2SJf1Jb/AMdoXxBpqjC/aAB2FnL/APE0v/CRaf63P/gHL/8AE0AJjXp/vNYWQ9FDzn8zsH6Uv9kXEv8Ax96teSDusZWIf+OgH9aP+Ei0/wBbn/wDl/8AiaP+Ei0/1uf/AADl/wDiaAHJ4f0xWDvbCZ/78zGRvzbNXo7eGEYiiRP91QKz/wDhItP9bn/wDl/+Jo/4SLT/AFuf/AOX/wCJoA0ZRmFx/smvHZdJnMrfvbTqf+XuP/4qvT5fEOnmFwDc8qf+XSX/AOJryKQ5lYj1NfK8QuK9nzLv+h9PkCk/acr7fqW/7Jn/AOe1p/4Fxf8AxVH9kz/89rT/AMC4v/iqo0V8rzU+z+//AIB9Rap3X3f8Evf2TP8A89rT/wAC4v8A4qj+yZ/+e1p/4Fxf/FVRoo5qfZ/f/wAALVO6+7/gndeAbKS21G6aR4WBiA/dzI56/wCyTXeHpXnfw5/5Cl3/ANch/OvRD0r7zJbfU427v8z4fOL/AFuV/L8j5Wv/APkI3H/XVv51Xqxf/wDIRuP+urfzqvX6Utj8xe4UUUUxBXqXwS/4/tX/AOucX82ry2vUvgl/x/av/wBc4v5tXHjv93l/XU7cB/vMf66Hr9FFFfMH1YUUUUAcB8RIZJL60McbPiM52jPeuO+y3H/PCT/vg17hiua1X4g+FdE1c6ZqerRQ3a7fMTy3YRbum9gCFz7kV89i8kWJrSq89r+X/BPfwucvD0Y0uS9vP/gHmn2W4/54Sf8AfBo+y3H/ADwk/wC+DXtM13bW9jJeTSKtvHGZWk6gIBknj2pbS5t76zhu7SRZYJ41kikXoykZBH1Brl/1cj/z8/D/AIJ0/wCsL/59/j/wDxX7Lcf88JP++DUtpbTi8hzDJ98fwH1r2CDUrK61C6sYJ0e6s9nnxAcx7hlc/UVbwKa4dinf2n4f8ET4gbVvZ/j/AMARP9Wv0FZHi2SaLwlqL2rOsywkoYyQwPtirsmqWUWqw6bJcKt5PG0kUPOWVcbj+GR+dLqWpWekadNfalOsFrCu6SVuij1NfVxfK0+x8tNc0Wu585/2v4n/AOf3VP8Av5JR/a/if/n91T/v5JX0opR1DLggjINR29zbXaM1rNFMqOUYxsGAYdQcdx6V6v8AaMf+fa/r5Hk/2bL/AJ+P+vmfN/8Aa/if/n91T/v5JR/a/if/AJ/dU/7+SV9K4HoKMD0FH9ox/wCfa/r5B/Zsv+fj/r5nzV/a/if/AJ/dU/7+SV7B8LJ7268Kyvqkk8sv2lgDOSWxgetdhBc210Zfs00U3kuY5PLYNsYdVOOh5HFTdOlYV8Wq0OVRsdGHwboz5nO4xoImGGjQ/Vaqy6Npk+fOsLaTP96JT/SrtFcB6Bl/8I3pQ/1doIf+uLNH/wCgkUn9gQL/AKi81CL6XsjfoxIrVooAyv7IvE/1Ou3w9nSFx+qZ/Wj7Jrcf+r1S2f8A67WZP/oLitWigDLzryfw6fP+Lx5/Rqa9zquMT6RDKO4iuQ3/AKEorWooA5Weygy0g0C+sZSOZbOVUJ9/kfn8RXllzf3IupR/bfiRcOePs+cc/wDXWvfaTA9BXTQrqjfS9/67HLiMO61tbW/ruj5//tC5/wCg54k/8Bv/ALdR/aFz/wBBzxJ/4Df/AG6voDaPQflRtHoPyrq+vr+X8v8AI5f7Pf8AP+f+Z8//ANoXP/Qc8Sf+A3/26j+0Ln/oOeJP/Ab/AO3V9AbR6D8qNo9B+VH19fy/l/kH9nv+f8/8zzj4Y3Ms+pXol1DVLoCIYW+i2KOeo+dua9IowB0FFcNap7SfNax3Uafsoct7nkXiS3mbxJfFYZCDKcEKazPstx/zwk/74Ne4YHpRgelfJ1OH1Um5+03d9v8Agn1dPPnCCh7PZW3/AOAeH/Zbj/nhJ/3waPstx/zwk/74Ne4YHpRgelZ/6uR/5+fh/wAEv/WF/wDPv8f+AeH/AGW4/wCeEn/fBrovA0EsfiaMvE6jY3JUjtXp2B6UYrfD5CqNWNT2l7O+3/BMa+eOtSlT5LXVt/8AgBRRRX0h86FFFFABRRRQAUUUUAFFFFAHjvxlt55tcsDDDJIBAclVJxzXnH2C7/59Zv8Av2a+qSAeormL74ieEtN1ptKvdYhiukcRuCjFI2PRWcDap+pFepRzB0qahy7eZ5VbLlVqOfNa/kfPv2C7/wCfWb/v2aPsF3/z6zf9+zX1BfXtrptjNeX0iw28K75JGHCj1qaJ45oUliKsjqGVh3B6Gtf7Uf8AL+Jj/ZK/n/A+WfsF3/z6zf8Afs1e0Oyul16xLW0wAuEyTGf7wr6QtdRsb27urW1mSSazcJOgHMbEZAP4VawPQUnmbaty/iNZUk78/wCAo6VzvjhGk8MyqiljvXgDPetf+1LL+1hpnnr9s8rzfJ5zszjP50alqdlpFi95qU629uhAaRs4GTgfqa8HEUvbUpU72urH0FCr7GrGpa9nc8a+y3H/ADwk/wC+DR9luP8AnhJ/3wa9v4I7VFa3VtewCazniuIizKJImDLkEgjI9CCPqK+a/wBXI/8APz8P+CfRf6wv/n3+P/APFfstx/zwk/74NH2W4/54Sf8AfBr2/A9BRgego/1cj/z8/D/gh/rC/wDn3+P/AADxD7Lcf88JP++DXq/hJCvhayV1KsFOQR/tGtO2ura8jMlpNFOisULRsGAYHBGR3BqavRy/KlgqjqKd7q23/BODH5o8ZTUHG1nfcMCjAoor2jxgwKMCiigAwKMCiigAwKMCiigAwKMCiigBkw/cvx/Ca8UltZ/Of9xJ94/wGvbqMD0rysxy1Y7lvK1r9L7nqZfmDwXNaN726nh/2W4/54Sf98Gj7Lcf88JP++DXuGB6UYHpXk/6uR/5+fh/wT1P9YX/AM+/x/4B4f8AZbj/AJ4Sf98Gj7Lcf88JP++DXuGB6UYHpR/q5H/n5+H/AAQ/1hf/AD7/AB/4B598PIZI9TujJGyAxDG5cd69BPSjFFe/gsL9Uoqle9jw8ZivrVZ1bWufLt9Y3Z1C4ItZsea3/LM+tQfYLv8A59Zv+/Zr6owPQUbR6D8q+kWaO3w/ifMvKk38f4Hyv9gu/wDn1m/79mj7Bd/8+s3/AH7NfVG0eg/KjaPQflT/ALUf8v4i/slfz/gfK/2C7/59Zv8Av2a9O+C1vNDe6t50Tx5jixvUjPLV61tHoPypcAdBWNbHurTcOW1/M2oZcqNRT5r28gooorzD1QooooAD0rxhdSg8NW3jfQ9b067n1HVdRubi02WzSC6jlA8sBgMfL0wemK9npCoJyRzQB4rpOi6/Amo2mpajq0X9j+HoHhgguHSNpzHIDwPvEYHHrj2rH1rXdZtNEaXVNV1y1uovDtpJYfZpZFVpSg81pMdXz/er6DxXO614F0PxBqP23U4JJJCgRwsrKsig5AZQcEZ9aAPOdVvfET3Gu/ZL3Uot11oywyRO/wAqsF80r2wcnd2Peo9RvdU0n+1tImu9clsl1hIred72ZTHGYgx3zANJs3f3T14zXtSRpHGqIoVVGAB2FKVB6igDxzwA+q3nibwrc6t9smmhstThkmugzOALhQgZm5J2gYJ5NUvik9/d3niizvbnVgfsUf8AZdral/JlH8eVHysfXPTtXuOBSYBOSKAPEte1bVv+E4xYz6vbNb39rAYzdS7GiZV3MsKr5ezk5ZiWzWdZzPo+j3GmifW45J9duBOReTQxxryVLOqlwG/2MZOM179tGc45o2r6CgD54n1bxMvhnQ7i71PVmmjhuEe2ee4glkZbllQ+YoId9gUYcEY571r3mqeKbjxth7u+sJ/tVp9iheaco0JCl1aJEKPn5gzMQQeeMV7gVB6ijAznHNAHnfwttF0/VfFdtJLfGf8AtaVylw7spQhdrjPGTzyPT2r0WgADoKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAPSvCzeJpHw98T+DdS0u7n16+lukgRbZnFy0pPluHxjAyDknjFe6Um0E5xzQB4nNpniGGw8TC61TWBNo+lWsdrFBcyLG0vkYdgB945H5mszxdr2rWNnr8t3q2t2mowWlqdNW2mkWERlU3swHy7slsluRxivoDFc3qvgLQda1VtQ1C2eSSQoZU81hHLsOV3LnBxQBwWqz6/L4l1W3hvNTggk121iVoZHXEJhG4Key59O/vWfeajq1jpcukXV1rbwpr13bwXLXs0e2JEygklUGRgSTtAIBI5OBXuGBQVB6igDyT4cPql34n0m71cXUk/9ivHLLcq24sJ2ADE98AdeaxPiC+oXuqeKbTULnVzdJcW39mWcJc2722ELNtHysd27JPIxxivd8YpMAnJHNAHit5qmqz/EeFbW51i2RtSNrPG13KVERjYZEYURqucEHk571haVcTaX4HsdJhm1yGdtQvfP23k8McLiSQopZQX+YFSACAWOT1r6H2jPSjavoKAPnq+1nxNFp2j3s+oatPOthETZrPPbytIHOSpUFJGPAIcGtx7/AMTXPxIuYp768s5TqKpBB58+w2pQcCJVMfPOXJyDXtO0elGBnOOaAPPPg9bLZaLqlq0l59oj1GfzYrl3ITMjbSu71GCcdc5r0SgADpRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB//Z)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mil3pkN44f1f"
      },
      "source": [
        "**1. Random Over/UnderSampling**    \n",
        "It literally increases and decreases the number of data randomly. When Undersampling by label, the group with the largest data (MCI) was used, and when Oversampling, it was applied to the minority group (AD). Since the data is randomly increased and decreased, the learning accuracy is not constant in the current experiment where the total data is small, and there is a possibility of generating outliers.         \n",
        "**2. TomekLinks**           \n",
        "I thought TomekLinks was the best resampling strategy for the data we had. The distribution of our data is so overlapping that when we learn, most of our data is predicted by the MCI group, which is the majority and median data, so the learning accuracy tends to decrease. Tomeklinks are pairs of very close instances, but of opposite classes. By removing the data of the majority class from each pair, it widens the space between the two classes and facilitates on classification task.       \n",
        "**3. SMOTE**      \n",
        "SMOTE(Synthetic Minority Oversampling TEchnique) is to create synthesizing elements in minority groups based on existing data. It works randomly picking a point from the minority class and computing the k-nearest neighbors for this point. The synthetic points are added between the chosen point and its neighbors. However, when the points of this SMOTE oversampling selects an outlier, we create data that is not related to the distribution of the data.\n",
        "\n",
        "We will compare the performance of resampling strategies through several classification models below and use the best one.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GH3RHQn-6Op0"
      },
      "source": [
        "![2.JPG](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAeAB4AAD/4RD6RXhpZgAATU0AKgAAAAgABAE7AAIAAAAQAAAISodpAAQAAAABAAAIWpydAAEAAAAgAAAQ0uocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEd5ZW9uZ055ZW9uIEtpbQAABZADAAIAAAAUAAAQqJAEAAIAAAAUAAAQvJKRAAIAAAADMjQAAJKSAAIAAAADMjQAAOocAAcAAAgMAAAInAAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADIwMjA6MTA6MzEgMTM6Mjc6MDgAMjAyMDoxMDozMSAxMzoyNzowOAAAAEcAeQBlAG8AbgBnAE4AeQBlAG8AbgAgAEsAaQBtAAAA/+ELImh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMjAtMTAtMzFUMTM6Mjc6MDguMjQzPC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPkd5ZW9uZ055ZW9uIEtpbTwvcmRmOmxpPjwvcmRmOlNlcT4NCgkJCTwvZGM6Y3JlYXRvcj48L3JkZjpEZXNjcmlwdGlvbj48L3JkZjpSREY+PC94OnhtcG1ldGE+DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgIDw/eHBhY2tldCBlbmQ9J3cnPz7/2wBDAAcFBQYFBAcGBQYIBwcIChELCgkJChUPEAwRGBUaGRgVGBcbHichGx0lHRcYIi4iJSgpKywrGiAvMy8qMicqKyr/2wBDAQcICAoJChQLCxQqHBgcKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKir/wAARCAGgAkMDASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwD6RooooAKKKKACiiq1zqVlZSIl3dwwM/3VkcKT+dNJvYTaW5ZooBBAIOQehrC8WeKrTwlpa3l3G8zSPsjiQ4LH69qcYynLljuKU4wi5Sehu0VznhDxnZ+L7WaS2ie3lgYCSJzkgHoc10dOcJQlyyWooTjUjzRd0FFUNZ1uw0CwN5qs4ghDBc4JJJ7ADk1HofiLTPEdq1xpFyJ0RtrfKVKn3B5pckuXmtoHPHm5b6mnRRRUlhRVVtSsVvBaNdwi4PSIyDd+VWqdmtxJp7BRRRSGFFFFABRRRQAUUjHCkgZIHQd6wbPxXBcXVzFeWVxp6W0nlyS3TRBd+Ado2uT0OelAG/RVGXXNLglMc2oWyOBkq0oBA/yRTLvXtPtI5Wa4SRoSu+ONgWXJAHH4igDRorPv9atLC1iuGJljlYqhiwckAn1/2TUc3iLTIY2Y3UbOu3dEjAuu4gcjPqRQBqUVVTVLGS4SBLuFpZM7EDglsZ6fkfyq1QAUVVXUrF7w2qXcLXA6xCQbh+FWqbTW4k09gori/F3xIs/CuqJYG0ku59geQKwUID0/Gum0bVrfXNIt9RsiTDOu4A9R6g1pKlOMFNrRmca1OU3BPVF6iisPXPGWh+HLmO31a88qWQbgioWIHqcDiojGU3aKuXKcYK8nZG5RUNpdwX9nFdWkiywyqGR1PBFTVOxSd9UFFBIUEk4A6k1WtdRsr5nWzu4Z2T7wjcNj8qdnuK62LNFFFIYUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUEgAknAHU1WttSsr2RktLuGd0+8sbhiPyp2YrpaFmiiuA1j4sabpPiF9NFpLPHC/lyzqwAB74HfFaU6U6rtBXM6laFJXm7Hf0UyGVJ4UliO5JFDKfUGn1kahRXPah478PaXq39m3t+EuQQrKEYhSexIGBXQKwZQynIIyCO9VKEopOS3IjOMm1F3sLRRTJZY4ImlmdY41GWZjgCpLH0VBa3ttfRmSzuI50BwWjYMM/hU9FmtxJp7BRRRQMKKKKACiiigAooooAKKKKACiiigArxz4jeDPEOq+L3vLK2kvLeZVERQ/6rAwQfTnmvYZJEijaSV1RFGWZjgAVBZ6jZ6grGxuobgLw3lOGx+VdFCtOjLnijmxFGFePJJlbw9Z3On+HbG0vn33EMKrI2c849aqeLPCtr4s0pbS6domjffFKnVT06dxW7Xn3xE8f33ha/t7LS4ImkdPMd5gSMZwAAPpRRVSpV/d77hWdKnRtU+HY3vB/g208IWcsdvK088xBklYYzjoAO1dHWB4L8QyeJ/DMOoTxCKYsySBfukjuPatSHVdPuLpraC9gknXOY1kBYY9qmr7RzfPv1LpezVOPJouhgeP/Cc3izRIre0mWKeCXzED/dbjBBqr8O/Bdz4Stbpr6ZJJ7kjKxnKqB7+tdpRT9vNUvZdBfV6bq+16hRXkWt/FvU9P8UXFrb2kBsreYxlXU72AOCc54r1qGUTW8cqggOoYA9siirQnSScuoqWIp1m1DoeG33gHxTJ42kkSCRt9z5i3gb5QM5Bz7DtXucalY1V23MAAW9TTqKqtiJVklLoKhho0HJxe4UVx+sfEzQtG1ttMuPOeRGCySRqCsZ9DzXXRSpNCksTBkdQykdwaylTnBJyVrmsasJtqLvYdRRRWZoFFFFABXOXPheV76e8t7mLzZbhplWaLcgDRCMgjPJ4zXR0UAcnH4FhjtEgFyH2BgHeMEnMPlDP5Zqp/wrqNkuYZLzdFNvwx37xvdXYfewBxgYHp6V29FAHJ6rpdt/ZWn6He6iIpAHEUrR4DjaygHtnDD64qO48AwXVtHDJc/KDLvKptLhynGQcj7n61v63afabHzEiWWS3PmKjDO8d1/EZqCCOe2t47nSH+0Wkihvs0jcgH+4x6fQ0AZtp4QntdVsboXybbTC5WMq8iAMFVudp4Yc4zwcda6iRS8Tqp2sVIB9DUFnqEF7uEZKyJ9+JxtZfqKwdQ8UzCWRdKjtBDE5je7vpvKiLjqq45Yj16UX1JlJJanmej+AvFNv43glmgkURXIkkvC3ysAck5759K9zrA0zxFNJcw22rQQxNcf8e9xbS+ZDMfQN2Psa366K+IlXaclsc+GoQop8j3OM8W/Dix8Vaml+1zJaz7QkhVdwcDp9DXTaPpNtoekwadZAiGBcLuOSfUmvL9a+LeqWHii4tba0tzZ283llXB3sAeec8V6sl5CbBLyVhFE0YkLSHG0EZ5rStGvCnGNR6dCKE8POpKVNa9SevNvHnw2vfEWvDU9MuYwZFCSJKSNuO4Neg2l/aahGXsbmK4VTgmJw2D+FWKxpVZ0J80dzerShXhyy2Mrw1oo8PeHrXTRKZTCvzP6knJrVqnq+oDStGu78oZBbxNJtHfArzjwb8UNT1zxTDp2o20AhudwQxKQUOMjPPIqo0alZSqLpuRKtSoONJ9djv/ABJZXOpeGr+zsH2XE0JWM5xz6Z9+leY/Dfwb4h0vxYt5e20llbRKwk3nHm5GAAO/PNexUU6eIlTpyprZhUw0alSNRvVBRTJpo7eB5pmCRxqWZj0AHU1yWi/EzQ9b1oabbedHI5IjeRQFkI9OayjTnNNxWxrKrCDSk7NnYUUUVmaBRRRQAUUUUAFFFFABRRRQAUUUUAZviKyudR8N39pYv5dxNAyRtnHJHrXl3w58GeINL8XpeX1rJZ28KsJC5/1mR0HrXsdFdNPESp05U1szmqYaNSpGo3qgrgdY+E+mar4hbUhdSwRzP5k0CgHce+D2zXdXEwt7aWZgSI0LkDvgZryfRfi3qV94pgtbq0g+x3E3lBYwd65OAc55qsNGu+aVLpuRipUFyxrLfY9ZhiSCBIYhtSNQqj0Ap9Q3V5bWMPm3k8cEecb5GCj9aW3uYLuETWsyTRN0eNgQfxrls7XOu6vY8t1/4S3upeKZ7y0vIhaXUpkcyZ3Jk5Ix3r1K1gFrZw26ksIo1QE9TgYqWsHxp4hk8M+GZ9QgiEsoISNW6ZJ6n2rolVq4jlpv5HNGlSw/NUXqzerk/iPouo654Te20nLSrIHaIHHmKO3/ANasr4deP77xTqFzY6nBEsscfmpJCpAxkAg5+or0Gk4zw1VX3Q1KGKpO2zPNvhP4Z1nQ2vp9Vhe1imVVSFzySD97HavSaKp6tqtromlzahfvsghXLEDJPoB71NWpKvU5mtWVSpxoU+VPRFyiuY8L+PdJ8V3UttYiaKaNd2yZQNw9Rg109ROEoPlkrM0hUjUjzQd0FFFFQWFFFFABRRRQAVmeI7iW08O3k9s5jljjyrDsc1p0yaGO4haKdFkjYYZWGQazqxc6coxdm0zSlJQqRlJXSaPO/C3iDVL3xJbW91eSSRPu3Kcc/KTXo9U4NJ0+2mWW3s4YpF6MqAEVcrjwGGq4ak4VZ8zvv/w5147EUsRUU6UeVW/rY53x3pV9rXg+7stLb/SGwQucbwDkr+Ncb8KvCut6Nq93eanbyWkDReWI5OC7ZznHt612vjO5ntfDkkttK8UgdQGQ4PWud8DanfXmuvHdXU0yCEna7kjORW0849hVWB5fi6/16GEcn9vTeO5vh6f16noNY+veFdI8SLF/a1t5rRfcdSVYD0yO1bFFdkZSi7xdmcsoxmrSV0UIdJgsNDbTtKRbaMRMkYX+EkdfrmvHvCvgTxLZ+N7aa4tZLeO3m3yXDH5WHfB75rqPiF8Q9R8NazFp2lww7vLEkkkyls56ADIrr/CeuN4i8M2upSxeVJKCHUdMg4OPau6Mq1Ck520kefONDEVlBPWBs0VV1O9GnaXc3hQuIImk2jvgZrzHwn8U9U1fxVbWGoW8H2e6fYnlKQyHtznmuanh51IuUdkdVTEU6Uowluzt7zwJ4evtZ/tO5sFa4LBm+YhWPqR3roQMDA4FFFZSnKVlJ3saxhGLbirXKOoa3puksi6lew2xk+6JHAzVyORJY1kiYOjDKspyCK8r+JHgXXNc8SLf6XH9qheNUKGQL5RH1PSu/wDCulz6L4YsrC8k8yaGPDkHIB9B9K3qUqcaUZxldvoYU6tSVWUJRsl1OO1/4SRax4kl1GHUPIhuJPMmjKZOT1wa9DtbdLS0it4vuRIEXPoBipaZMzJC7RjcwUlR6nFROtUqJRk9jSnRp0m5RW4+ivCfD/jLxTP44t45bq4m8252S2rZ2qpODx2wP5V7tV4jDyoNKT3Iw+IjiE3FWsFFFFcx0hRRRQAVz/jO+udP0HzrKVoZPNVdy+nNdBUVzawXkXl3USSpnO1xkZrDEU5VaUoQdm1ub4epGnVjOauk9jifBOtajqOryxXt08yCIkK2Oua6jT/9Dv7jT24T/XQD/ZJ5H4H+YqzbaZZWche0tYoXIwWRADioNXjZYY72EZltG38d06MPy/kKxwVCph6PJUlzPua42vTr1uenHlXYr+KJksdDub6ONftKR+VFJjlS5Cjn6kVwnxF8GapdabpEei273MNnEY5IkPO44+bHfPNdv4tj+2+D7p4PnAVJ1x3COr/yFc9421e6jm0+XT7qSKOaDf8Au3IBzjFa1sf/AGevrFr2/XQ5qWB/tCp9Xva/6akfhzwzqOn/AAyntdTDRXaM1zboTkwsOV+hyP1ruNMuxf6Va3Y/5bRK/wCYrk9D1Wb/AIQXU7u+neVow4Bdsn7vA/M10+h2rWWgWNvJ9+OBFYehxzVUsT9bSr2tzak1MN9UqOhe/LoZ154G8P3+sf2nc6er3G4M3JCufUjoarfEPRr/AFvwfNaaTzKHVzEDjzFH8NdTXkF58XNUh8VSW8VpD9hjuDEY2U+YQDgnOev4V6WHjWqyTjry9zhxEqFKDjPTm7Gj8KPC+s6LdXt1qkElpDIgRYZOCxz1xXp1Ijb41YDG4Zwa5nx74nn8K+Hhd2cSyTySCNN4yq57monKeJq7as0hGGFo2vojpZI0mjaOVQ6OCGVhkEelYekeC9C0PUXvtOsljnbOGLE7Aeu3PSsP4ceN73xYLyHUoY1mtgrCSJSFYHPGPXiu6qJxqUZOm2VTlTrxVRK/YKzv+Eg0gal/Z/8AaNv9r3bfJ8wbs+laJGQR0rxIfDDxGPGHmEKbf7T5v2zzByM5zjOc1dClTqX55WsRiKtSny+zje57Ne2kd/Yz2k4zFPG0b49CMVwHh74TRaL4jj1Ga/8AtEUDb4o9mDntk16MBhQOtFRCtUpxcYvRmlShTqSUpLVBVTUNVsNKhWXUruK2RjgNI2MmrdedfFHwhq3iOSyuNIQT+QrI8O8L1P3hniijCM6ijN2QV5zp03KCuz0C1uoL22S4tJkmhcZV0OQalrlvh54fvfDnhZbTU2HnNK0nlhsiMHtXU1FSMYzai7oqnKUoKUlZhRRRUGgUUUUAY/im7nsvDtxPayGOVcYYdua5Xwfrup3/AIhjgvLt5YijEq2OuK76eCK5hMVxGskbdVYZBqC30qwtJhLbWkMUgGAyIAa8vEYStUxUK0Z2irXWuup6dDFUaeGnSlC8nez00LdFFFeoeYFFea+MNV1C18SzxW15NFGFXCo5AHFdp4XmluPDVnLO7SSMmWZjknmvNw+YQr4idBRs43/B2PRxGAlQw8K7d1K34q5rEAjBGQawLLwP4f0/WTqdrYIlzkspySqE9wOgrfqC+uRZafcXJUuIY2k2jqcDOK9WMpLSL3PLlGD1ktjhviv4d1bXdOsm0mNrhbd2MkCHls4wcd8YP51c+F+hanoXhqSLVkaJpZi8cLHlBj9M1zPhj4q6rq3iu2sb22g+zXUmxRGpDJnpznmvWq7K7q0aSoTStucVBUa1V4iDd9gqtqGn2uqWMlnfwrNBKMMjd6474j+N73woLSHTYY2muMsZJVJUAdsetafgLxRN4r8Pm7u4VinikMb7B8rHrkfnWHsKkaardDo9vTlVdHqaGg+F9J8NxyLpNqIjKfncncx9sntWuSACScAdTRVPV7SS/wBGu7SGQxSTwsiuP4SRjNZOTnK8maqKhG0UR2Wu6XqN09tY38E8yfeSNwSKj8R6HD4j0G40y4cxrMBh16qQcg/mK8x8EfDvxBpXi+C8v41tre2YkusgPm8YwMH+dexVvWhCjUXspXMKE516b9rGxw/gn4cp4U1KW+mvftUzIY0CptCg9fxrt3bZGzYztGcDvS0VjUqzqy5pu7NqdKFKPLBWR4zZfFfW7jxVFFJDF9jkuBH9nCfMFJx19a9mrITwrokesHVE02AXhOfNx39cdM+9a9a4ipSm17ONjHD06tNP2krhRRRXMdQUUUUAFFFFAHN+J/FEvh+4gjitkm81SSWcjFWvDOuya/YyzywLCUk2YVs54pdc8M2uvSxSXMsqGIEDZjmp9E0ODQrV4LaR3V33EvivLhDG/XHKT/ddtO337npylg/qajFfvPn3+4vywxzx7Jo1kT+6wyKjhsra3ffBbxxtjGUQA1wXjf4lz+GtcGm2FnHMyIGleUkdegGK7Lw9rCa/4ftNTjjMYuEyUP8ACehH5ivYnhpRiqsluePDExlJ0ovVbmlRRUZuIVYhpUBHUFhWTaW5sk3sZmteFtH8QvG+rWazPHwr5IOPTI7Vo2lpBYWkdtZxLDDENqIowAK8P8bzeJz48n8trwYkH2TyS20r2xjivbtPNwdNtjef8fHlL5v+9jmuytSlTpxbldPp2OKhVjUqTSjZrr3J3RZI2SRQysMEEcEViaZ4N0HSNSa+sNPjiuDnDZJ259AelWPEHiLT/DOm/bdUdljLBFVFyzH0AqLw54q0zxTayT6U7nym2uki7WX8Kyiqqg5K/KbSdJ1FF25jH+J99qth4RMmjmRCZQs0kY+ZE55Hpzisb4QalrF/b366lLNPaoV8qSYkndzkAn8K9JZQylWAIPUEdaRI0iXbEiovoowKtV0qLpcuvczdBuuqvNp2HUUVXnv7O1lSK5uoYpH+6ruAT9Aa50m9jpbS3H3V1DZWktzcuI4YlLu56ACsTQvG2h+Irx7XTLovMq7tjoVyPUZ61oa9pS65oN3pryGMXMZTeP4feuG8D/DO78O+IP7T1G7ifylZYkiz82eMnP8AKuinCi6UnN2l0OapOsqsVBXj1PQUsLSO6a5jtoVnbrIEAY/jVikZlRcuQo9ScUKwZcqQR6g1z6vc6VZbC0UUUhhRRRQAUUUUAIx2qT6DNcH/AMLEuGuPJOnREF9h/eH1x6V3hGVI9RXKD4facJ/N+03Gd27GR659K8vMIY2Tj9Vdu+36np4CWDjzfWlftv8Aoa+mAQNPpco3Ig3xA94m7fgcj6YrH8mbQ7hrRtPXVrEHNsEKmaAH+AhjyPQ1oa9cy/abXTdOPl3l0DuuAMmGFeWYe/QD3NeUJ4u0f+3vskuhRvamfyjdNKxuOuN+7PXvXrQw88RdQV7Hh1sTGjJNux6bBY3etXUX2rT/AOzNLgcSfZmxvncdCwHAUeneunrC0W4ns9RuNEv5mmeFBLbTP96SI8c+pU8flS+ItSuYmt9M0oqt9ekgSN0gjH3pD9O3vWMbRib36vcm1HxLp+nXQtS0lxdkZ+z2yGRx9QOn41zM9v4aGsDV9V8OXlnIXDG5nh/dhvVgCQPqa3odOt9G0G7i0Rke+aF2ErMC80uOCT3Oa8k8F/8ACQz+NEiuxdyRSFlvVuNxQoQd24HiuzD0p1ISnGdrHHianJOMJxvf8D3hWV0DIQysMgjuKq6npdnrFi9nqUCzwP1VvX1rM8FMzeFLXLFkVnWInugchf0qDxF490TwxepaajJM07LuKQpu2jtnnisaSnNrkWp0znBQvPY09F8PaZ4ft3h0m1WBXOXIOS31JrzX4uarrllrVrFaT3FtZeVlWhJUM+eckV6np2oW2q6fDe2MgkgmXcjD0qaSKOYASxq4ByAy5relWdOrzzV/Uyq0VVpckHZeRgeA7nUbvwbZTaxu+0sD8zj5mXPBP4V0VAGBgUVjOXNJyta5vCPLFRvewVl694i07w3Yi61Wby0ZtqgDJY+wq5HqFnNctbxXULzL96NZAWH4VzXj/wAHSeL9MgjtbhYbi2cum8fK2Rgg1VKMXUSqaIirKapt09WbOg+ItO8SWJutKm8xFbawYYZT7itSuS8AeDpPCOmzpdXCzXFw4Z9g+VcDgD1rrScAk8AUVVBTapu6HRc5U06isworl7f4i+HbrXBpUV2xmL+WrlMIzegauoqZU5Q+JWKhUhP4XcKKKKgsKKKKACq2pXZsdMuLpVDmGMuFJxnFWagvbVL2ymtpCQkqFCR1ANRU5uR8u/Quny865tjktH8dT6nq9vZvYxxrM2CwkJxx9K7Sua07wRY6bqEN3DPOzxHIDEYNdLXDl8cXGm1ind38tvkduPlhZVE8KrK3nv8AMglsLSeQvNbRSOerMgJqWONIYwkSKiL0VRgCnUV3qMU7pHC5Sas2FIyhlKsMgjBB70M6ouXYKPUnFNSaJ2wkisfQNmndXFZ2MXT/AAZoOl6odQsdPjjuMkhskhc+g7Vu0UVcpSk7ydyIwjBWirGbrXh7TPEFukOrWqzqhypJwV+hFT6Zpdno9ilnp0CwQJ0Vf51z2p/Enw9pOtNplzNMZkYK7xx7kQ+hOa6pHWSNXjIZWAKkdxVzVWMUpXt0M4SpSm3G11ueHePda8SW/jq4iiubuBFdRbRxEhWXHBAHWvaNKe4l0e0e+XbctCplHo2Of1qw0MburvGjMv3WKgkU+tK1dVIRio2sZ0aDpzlJyvcKKQkKpLEAAZJPaoLa/tLwsLS5hnKfeEbhsflXPZnTdbGXr/i/R/DLxJqtyUklGVRFLHHrgVpabqVrq2nxXthKJYJRlWFcT8Qfh5c+KdRh1DTrmOOZYxE6S5AIBJBBH1rpvCXh4eGPDkGm+d5zIS7vjALE5OPauicKKoqUX73VHNCdZ1pRkvd6M2qK4T4rarq2leH7dtIklgWSbE00X3lGOBntmofhNrGsarpN3/a0ktxFHIPJml5Jz1Ge9H1eXsfbX0D6zH2/sbanoNFFFcx1BRRRQAVyl948tLG+mtXtJWaJypIIwa6uuK1LwFJf6lPdC+VBK5bbs6Zrzcwli1BfVVrfXb9T0cBHCOb+tPTpv+h2FrOLq0inUFRIgYA9sipahs4PstlDAW3GJAmfXAqavQjflXNucErcz5djkPFXhnwrreoxy63cLb3SKASs4RmXsCDXR6THYQaXDBpBjNpEuyPym3AAe9cZ4x0LU9Q14z2do8sflqNykda6PwhZXGn+H44LyIxSh2JU/WuClj8TVxMsPOLUI3s9f+GO6rgcNSw6xEJLnla60/4c3K8i8Q2903iK/KRTFTM2CFOK9dopZhgVjYKDlazuPAY54Obmo3uinpSf8Sey3r8wgTqOR8oq5RRXoRjyxUexwSlzScu5zfjfwkvi7R0thP5E0MnmRORkZxgg1W8CeCR4QtLjzbgXFzcEb2UYUAdAK62iuj21T2fsr6HP7Cn7T2ttQorzPVvjAmneIprGPTfNtoJfLeYyYY4PJAxXpMUizQpKnKuoYfQ0VKNSkk5rcKdenVbUHew+vFPiB4Q8R3/jSW5trWa8huGHkSR8iMY6H+7ivVPEviSz8L6Sb6/3Mu4IiJ95ie1VPCfjTT/F0c/2JZIpYCN8UnXB6H6Vth5VaKdWMbrYwxMKVdqjKVnuaui21xZ6HZW99J5lxFAiSvnOWA5q7RXiuufFHX7TxdPFbbIra3n8sWzRg7gDjk9cmoo0J4iT5TStXhhormOn+L1nq91otodMWaS3SQm4SHJPT5SQO3Wp/hNaaraeGphqqSxxvNm3SbIYLjnr2zXcW8pntYpSpQyIGKnqMjOKkoeIfsfY2EsOvb+3v8gooormOoKKKKACoru4FpZy3DAsIkLEDvipagvbc3djNbhtplQrn0yKmfNyvl3KhbmXNsc5p/ju11DUILRLSVGmcKGJGBXVVxel+ApNP1S3uzeq4hcNtCYziu0rz8vli3B/WlrfTbb5Hfj44RTX1V6W13/UwIPn8fXhf/lnYxqmfQsSf5Vhf8Iz4Gj8RHUjPEJxJv8AJMw8sPnrt+tbesN/ZHiC21iQEWskRtblx0jycox9s8Z964m48MaxLqEksdi7RtKWVgRgjPXrSxmOxODS+rxve99/0M8DgcNi3JYiVuXVbfqdtqpC+MdBlj+86zxkjupUH+YFVLxDJ4o1vcCZF0pBCO+CXzj8cVPpxGseJhfpzaabAbeN+zytjeR7AADP1qr4k1TQmvI5U1oWep2+Qk1uvm4HdXAyCPY11yXNB36/5HApckubszlPD9nep4hsmkgnVRKMllOBXc+ItSaR/wCxNIw+pXg2uVGfIjPV29OOnqaoabNqviOFvs3ie1MKnEhs7bbKPb5idv1xXQ6Votno0LJaIxeQ5lmkbc8h9WPeuPAYL6pTcIyumzvx+OljpKXLayJ7Cyi07T4LO3GI4Iwi/QCuG8a/DNvE2tDUrK9W3kdQsqyKSDjoRXoNFexSqzoy5oHm1aMKseSa0M3w/osXh/QrbTYHaRYVwXbqx7mtKqup38Wl6XcX0+THbxl2A6nFcL4V+Kg8Q+Io9MuNOFsJ8iKRZN3I5wRirjSq1VKole25EqtKi4027X2PRKiu45JbOaOB9kjoQjehxwalrK8R+IbPwxpDahf7igYIqJ952PYVlFOUko7m0pKMW5bHkfhfwT4ntPHVtNPbTQLb3G+a5Y/K6g84P8Wa9xrnPCfjbT/FyziySSGWDG6OTrg9xWrrV++l6HeX0cfmPbwtIqepArqxNSpVqKM1ZrQ5MLTpUqblTd09S9TZEEkbI3RgQa8i8DfEbXdV8XQWGpMlzBdEjasYUx8ZyMduO9ev1lWoToS5ZGtCvCvHmieVWHwdltPEkVy+oI1jDKJFUKd5AOQD2r1WiilVr1KzTm9h0aFOimoLcKKKKxNwooooAo6vqaaPpr3kqNIqEAqvU5OKy9E8YW+t6h9khtpI22FtzEY4rR13SzrGkyWayCIuQdxGcYOaxvD3g59E1T7W12so2FdoXHWvLryxqxUFSX7vS+3z8z06EcG8LJ1H+81tudVTZJFijZ5GCooyxPYU6q2pRvNpdzHGu53iYKPU4r0ptqLaPOik5JMgi1/Sp5kihv4HkchVUOMk+laFeX6N4a1i31yymmsZEjjnRmYkcAH616hXn5fiq2JhKVaHK0/P9Tvx+Go4eaVGfMn6foc/42SR/DEwiVmbevCjJ61yngWG4TxIDLHKq+U3LKQK9Loqa+Xqtio4nmty209CqOYOlhZYflvzX19Qooor1DzDzLXPhENU8Sy31vfiK2uJDJLGy5YEnJxXpNtAtraxQR52RIEXPoBipKxvFXiGPwxoE2pSxGYoQqRg43MenNdEqlWvywevY540qVDmmtO5s0VxXgb4hr4uvLizmsxa3ESeYoV9wZcgHnHqRXa1nUpypS5ZrU0p1YVY80HoY/iywvNT8K39nprbbmWLCc4z7Z9+lecfDDwpr+l+KGvL62ls7ZI2SQSceYT0AHfnnNdJr3xV03Q9ek002k1wYWCyyoQAp7jHfFdraXUV9Zw3Vs2+KZA6N6gjIrpU6tCi4uOkjkcKNespqWsSaiuc8d6/c+HPCs99Yxh59wRWYZCZ/iNcx8MfG2r+ItQu7LVmWcRx+YkoQKV5xg4rKOHnKk6q2RvLEwjVVJ7s9IkiSaMpKiuh6qwyDSRRRwxhIY1jQdFUYAp9Fc50BRRRQAUUUUAFFFFABRSH7p+leST6/qy6lIg1C4CiUgDf2zXnY7MIYJRck3c9HA4CeMcuV2seuUUyIkwoTySorlPFPxG0zwtqS2NxDNcTlQ7iLHyA9M5716lOnKq7QV2eXUqQpK83ZFzxnql3pWkxTWMnlu0oUnGeMVn+Cdd1DVr65jv5vMVIwyjaBg5rdt30zxXottdGNbi1mAkQOOh/xqxY6PYaZIz2NqkLOMMV7ivJqYXE/XVV57QW6u/y2PVp4rDfU3S5Lyez0/Mu0VjeKdUuNI0Rrq02+YHVfmGRg1i+E/FGoazqzW955WwRlvlTBzWtTHUaeIjh5X5mZ08DVqUHiI25UdnRRRXccJx9/wDDLQNR106nNHKGd98kKthHPvXXqoRQqjCqMADtWH4q8WWPhKwjuL5XkaV9scUfVvX8qPCviyx8WafJc2KvG0T7JI5Oq+ldElWnTU5XcUc8HRhUcI2UmL4s8L2/ivR/sNzI0LK2+ORRna3071R8FeBrbwfHOy3DXNzPgPIV2gKOgArT8S+JLPwvpJv74My7giRp95ye1U/CXjXT/F0c/wBiSSGWAjfFJjOD0Ix2pp1/YtL4CWqHt038Z0dZdx4a0a61MahcabbyXQIPmsnJI6H3rUrnvFnjKw8I28L3qSTSTkhIo8ZIHU81lTU3Llhuzao4RjzVNkdDXDeKfifY+GtaOm/Y5bqSMAysrABc9vfiuh8M+JbPxTpP26wDoFco8b9UYdv1rB8T/DLTvEus/wBotcy20rgCUIAQ+OM+xxW1GNOFRxroxrSqzpKWHev6HWaZqEGq6Xb39qSYbhA6Z9DVqq2nWEGl6bBY2i7YYECID6CrNc0rXdtjpjey5twooopDCiiigAoorJ8TzSW/hu7lgdo5FThlOCOazq1FTpym+iuaUoOpUUF1djTmhjuIXhnRZI5FKsjDIYelYX/CH2ioYkvtQjtf+fdblggHp64rlPCGq39z4mt4ri8mljYNlWckH5TXYeMLySz8L3Rt32SzbYEf+6XIXP5E1x4PGQxlJ1VG1nY6Mfg3g6nJN30uZdtbnxCxstPLWXh60PlfuPla6YdQD2QevesrxDqM3hjVRY6GsVnbiJTsjjHJ56nvT9O+JXh7T9Qg0C1gmWCFhbrcADYSOM+uM967G90TTNQm+0XtpHK+3G9vStMwwWJlRUYPlk9b+XyMssxWGhVc6i50tP6ucfpk8sl5oWr5Au7u4ktrgqoHmpgkZA64x1rs7/WdN0sf8TC9gtyegdwD+XWuBW6l1zxmuneEkSCy0tWBusZSN34ZgO5xwPxqTXYovCWowfYoY7q5njLy3V4vmyOc46np9BXP7f6nhlKq72sm/PYuFN43FSjh1ZNtq52+n69peqyMmnX8Nw6jJVG5A+laFecy3L3/AIRm1/y4bfUdNmDwywoEyMjKnHUEE16Ijb41YjBIBx6V2YetGvTU47Mxq0p0asqU90MubeK7tZLe5QSRSqVdT3BrmdB+HWieHtXOo2iyvMM+WJGyI8+ldXXOeJvHOkeFZ4oNRaR5pV3COJckD1Nd1J1XeFO+pyVVSVqlS2nU6OsXxX4Zt/FeitYXMjREOJI5FGSrD279avaVqlrrWmQ39g++CYZUkYI9j71geK/iBpvhO8itbmGa4nkXeVix8o980U41PaWgveQ6s6Xs7zfusPBXgS28HrO63LXVzPgNIV2gKOwFdUyh1KuAykYII4IqjomsWuvaRBqNgxMMw4BHKnuD7ir9TVlOU257jpRhCCVPYzNP8OaPpV09zp2nwW80nDOi4P8A9atJmCIWY4CjJNcZ4h+J2k+HtabTZoZ55I8ea0eMIT25611ttcQ6hYxXEJ3wzxhlPqCKqpCokpz6ip1KTbhTtocJZ/F3TLvxAmnizmSCSXykuSwwTnAOPQ16FXB2fwm0e08QrqKzTNDHJ5iWxxtU5yOfSu8q8R7C69iZ4f29n7b5BRRRXMdQUUUUAFFFFABRXNeOb25sdFiks53hczAFkODjBrK8B6nfX2pXSXl1LOqxAgO2cHNebPMIQxawrTu+p6MMBOeFeJTVl0O6ooor0jzjzbXPFer2euXdvb3W2OOQhRtHArvdInkutGtJ5m3SSRKzH1OKguPDmkXVw81xYxPI5yzEck1oQwx28KQwqEjQYVR2FeXg8LiaNac6s7p7K701PTxeJw9WjCFKFmt3Za6D6K4XxJ4u1PS9cmtLXyvLQAjcmTyK6nQL6XUdDtrq42+ZIuW2jA61tRx1GtWlRhe8dzKtgatGjGtK1pbGjVHWdHs9d0uWw1GPfDIOcHBB7Ee9XqK702ndHA0pKzOd8L+CdK8KGV9PEjzSjDSynLbfT6V0VY2seLdE0C5S31W/SCVxkJgk49TgcVqwTxXVvHPbuskUihkdTkMD3rSp7SXvzvr1M6fs4+5C2nQ4jXvhVpuua8+pG7mt/ObdNEighj3IPau1tLWKxs4bW2XbFCgRF9ABgVx+u/FHR9D1x9NlhnnaIhZZIwMIe498V2VtcRXdrFcW7B4pUDow7gjIrSr7fkj7TboZUfYc8vZ79RLm2gvLZ7e6iSaGQYdHGQRVXS9E03RY3TS7OK2DnLbF5b8av9K4V/izoia8dOMU+wS+UbjA25zjPrjNRThVqJqGppUnSptOdkzsNS1CDStMuL67JENuhd8dcCuR8LfE6y8S62NNFnLayOCYmZgQ2O3tXX39jBqmnT2V2u+CdCjjPUGuU8MfDTTvDWsnUUuZbmVQREHAATP8zV0vYezlz/F0M6vt/aR9n8PU7SiiiuY6gooooAKKKKACqh0qwLbjZwE5znyxXNXXxBhtryW3NhIxjcpneOcHFdbDJ50CSAY3qGx6ZrjpYjD4puMHzW8v8zrq4fEYZKU1a/mPAwMCuH8YfDW28U6qmoR3jWkxULL8u4OB0Psa7jNFehTqTpS5oOzOCrShVjyzV0UdF0m30LR7fTrPcYoF2gt1PqTXOfEvxLf+GvD8UulgLNPN5ZlK58sYz+Zrc1/XE0GxS4khaYM+zapxWdpet2PjFbixu9OBhVAzLKQwPNc0cdQhilTqO8n07nRLA154V1KatHv2Of8AAGtXPjjR7yy8RJ9oW3dWEyjbuz2OPSuy03w5puk3JnsYCkhXbkuTxVrT9MstKtvs+m20dtFnO2NcZNWGkRDh3VT7nFaYinQqVvbKCT6GeHnXpUPYubfcr6lqdnpFk13qVwlvApALue/pTNK1jT9btPtOl3SXMQO0sh6H0Nc38RvDV34q8PxRaVIjTW83mCMtgScEYz681W+GXhDUPDFldyaqypLcsMQq2doHc+9dSp0nQ5+b3uxyupVVfk5fd7mv4z8H2/i/TooZZmgmgYtFKBnGeoI9OKPBvg+38IabJBFM1xNM26WUjGfQAeldHXJeLfiFp3hO8itJ4JrmeRd5WIgbR7k0oSrVI+xjquw6kaNOXtp6PuaPizwxb+K9GNjcSNCytvjkUZ2t9O4qj4K8DW/g+KdluGubm4wHkI2gKOgArb0TWbXX9Hg1GxJMMwyAw5Ujgg1dkkSGJpJWCIgLMxPAAqfaVYxdK+nYr2dKU1Wtr3HVy/jXwTb+MLaDdO1tcW5PlyAZGD1BFXdJ8Y6Frl81ppl+k06gnZgjIHpnrW3STqUZ32ZTVOvBrdGF4R8LW/hPRjZQStM7v5ksjDG5sY6dhxW7RRUSlKcnKW5cIRhFRjsgoooqSgooooAKKKKACmyxJNGY5UV0bqrDINQajeDT9Nnu2QuIULlQcZrndK8cw6pqcNmtk8ZlOAxcHFctbFUKU1SqOzlsjqpYWvVg6lNaLc6KHTrOCQSQWsUbjoyoARVbxBpI1vRJ7HzPKdwDHJj7jA5B/MVpUm4eo/OuhRilZKxzSbn8Wp4vb+A5bXxGLnUdM1F1SXzDFaoro5znh8jAz2xXe6iNY1exme/Q6TpcUZeSJX3TTKBnaSOFH05rrK5XxP4og0+abS57R5VlhwzK4HDAioxmLUIKeInotP6sThcA5ycKEdXqeY6l4l1bwilpY6MUslmhF1KwjBMjPz37AYH4GvTNEtrbxp4b07VdbtT9paLacMVB5POPfrWLpF7peoaDZxeJ9FkuGtQY4bgW5lV1BwOR0PqDWtqXiK+GkSnQ9OeytYIyWu7qPYqKB0ROpPp0Fazq4XEYWEFBP9TCjHEYTESk5tdLdUZviS60jSri08NWYciaZbi5hizI7KOQgHqxA/Cr2qeOrnSXj+3WFtZCUZjjurrEhHqQoOKk8D+G4dIszqWpSifV70eZNLKwLJnnb/jXOfEL4favr3iIalpLRzRyxqjI77ShH9KMHRpSlyyfKvIWJqYjldVK8md3oXiGLWfMiaI29zGqu0ZYMGRujqw6qfWuc8dfDo+K9Qhv7S8W3nVBG4kXKsB0PHeneHNHk0PXdH0wyiWe106U3Lr0AZxtH0znFdzWlOpKjUbpvbqN01Xp8tVGT4Z0GLw1oEGmQyGUR5LOf4mJyTWD40+HcHiy+hvEu2tbhF2Odu4Mv+NV7v4s6PaeIG00287okvlNcKRtBzgnHXGa7tWDKGU5BGQa2br0Z+0ejZCVCvB0lqkZ3h/RLbw7osGm2ZZo4gcs3VieSTWlVHV9ZsdDsTeapOsEIOMnnJ9AKZouv6b4htGudJuVnjVtrcYKn3FYtTknN/ebxcItU0/kcl4n+Flp4g19tSjvXtvOIM6BN24+o9OK7exs4tPsILO3BEUCBEz6AVPRTnWnOKjJ6IUKNOEnKK1YUUUVkahRRRQAUUUUAFFFZuu6wuiab9reIyjeF2g461FSpGnBzm7JF06cqk1CCu2Xp7eG5QJcRJKoOQHXNMgsrW1YtbW8cRIwSigZrE0DxbHr189tHavCVTfuZga6KsaNWjiI+1p6+ZrWp1sPL2VTTyCik3D1FLXSc4UVyWo+PIdO1Ge0axkcwttLBwM101jdC+sIbpVKCVA4UnpmuWji6FabhTldrc6auErUYKdSNk9jPvvC+laleNc3cBeV8ZO8itCzs4bC0S2tV2xRjCjOcVPTPOiJwJEz/vCtI0aVObnGKTfUiVarOKhKTaXQfRRRWxieYePfhvqev+I/7S0uaJlmVVkSVsbMcZHtXe+HdKOh+HbLTWl81reMKX9T1P4VfmmS3geaZtscalmJ7AVxWi/FTSNZ15NMSCeDzX2QzSY2uew46Zrr5q9elypXUTj5KFCrzN2civ4h+FFnrniF9SjvpLZJ23TxBc5Pcqe2a7qztYrGyhtbddsUKBEHoAMCpqy9a8S6T4eSNtXu1t/NPyAgkn8BWbqVayUHrbY1jTpUW5rS+5qHnrXnL/CCwfxIb830n2RpfNNvt5znON3pmu907UbTVbGO80+ZZ4JB8rr3qzShVqUW1F2HUpU6yTkrh0ooorE2CiiigAooooAKKKKAOMuvh8lzey3BvmXzJC+NnTJzXYQx+TBHHnOxQufXFPorloYSjh5OVKNm9zqr4utiEo1HdLY8x8aS3S+JphC8wXauAhOOldp4RZ28L2hlLF8Nkt1+8a2GjRjlkUn1IpJD5Nu7IANqkgVyYfAOhiZ4hzvzX07a3OqvjlXw8KCja1tflYzPEOhjXrFLdpjDtffuAzVXw54WXQLmaVbkzeagXBXGOc1gWvxAvp7+GBrSALJIEJBPGTiu8aaJHCPIiseilgCaWHeDxtV4imryj11HiFi8HS+r1HaMumg+vPviDBcS6zbGCOV1FvglFJ/iPpXoNFdWNwqxdF0m7HNg8U8LWVVK5zPgOOWPQGE6Ojec3Dgg101FFa4ej7CjGle9kZYit7erKpa1wrjfGPw7tfFl9Fem6e1nRNjFVDBh/jXZUV106kqcuaDszkqU4VY8s1dGdoGiW/h7RYNNsyzRwg/M3ViTkk1JrOn/ANraLd2HmGP7REY94/hzVXV/FWi6Fcx2+qX8cEsgyqHJOPU46VqxSpNCksLB43AZWU5BB70PnT9o+vUFyNezXToeW+C/hhqei+KI9R1K4hEVqSYxE2TIcEfgK9Upk00cELyzOqRoCzMxwAB3rL0jxTo2uzyQ6VfR3EkYyyjIOPXnrWlWpUrvnl0MqNOlh1yRe5ynxcn1mHRbQ6U0yWxkP2hoCQw/u5xzjrU/wmm1abw3OdWMzRib/R2mzuK45684zXdMoZSGAIPUEUABVAUAAdAKft17D2XL8xfV37f23N8haKKK5jqCiiigAooooAralZ/2hps9oX2echTdjOK5vSfAy6XqkN4L1pPKOduzGa62iuWrhKNapGpON2tjqpYutRpunB2T3Gyf6tvoa8bP2z+0P+W+PO/2v71ezU3y0/uj8q5cwy/65y+9y2OnAY/6nze7e5T1HWdP0lVOoXSQlvuqeWb6AcmuL1o6D4g1YXDawbUlAgEsDIOPdgBVoaiLWDWNeeCOa9jvmtVEoz5canAA9PX8aseHfE8viLUWsr2ztxF5ZbgE5/OpxVbDVaiwtbVvZa/mRhKeLhTeLoaJXN3w9p0WmaNHbQXC3MYJYSLjByc1V8ZozeFrggEqjxvIB/dDgn9BVa6sv7EeS+8OMGSP5rnTkbKuvcqP4W+nWt2zu7bVtNiubciW3uE3DI6g9iK9GFOMIKlFWSR59SpKtJynuzyfVYri41i6mtY5ZIZJWZHRSVIJ4Ir0aXWrXRdDtTdFnnaJVitkGZJWx0AqD/hD4YWYabqV/YQMc+RBL8i/QEHH0FXdM8Oadpcxnijea6b71xcOZJD+J6fhXnYPL3hak5qV+Y9DGZhPFU4U+W3L5kPh7TbqE3Gp6qAL++IZ0ByIUH3Ywfb+dbdFFeqlZWPOSsjz27+Eem3XiJtQ+2SpBJL5r24UcnOSM+ma9BVQihVGABgClrFi8X6FNrB0uPUoWvN2zyxnlvTPTNdEp1ay11sYxp0qL00uZ3xA8KXHizRI4LOZY54JPMQP91uMYqp8OfBV14TtrqTUZka4udoMcZyqAe/c8121UtV1iw0Sz+1apcpbw52hm7n0FONao6fsVsyZUKSqe3e6LteJfE6bxEvjIrG12tsFX7L5G4KeOenfNewaVrFhrdn9q0u5S4hztLL2PoRVwqGxkA46ZFOhVeHqXcbir0liKdlKxneHWvW8OWB1TP2swr5u7rnHf3rSoornk7ts6YqySCiiikMKKKKACs3XtHGt6b9kMpiG8NuAz0rSorOpTjVg4TWjLp1JU5qcN0c34e8JLoN+9yt0ZiybNpXHetTXd39gXuzO7yWxitCggEYIyKxhhadKi6NJWTv+JtPE1KtZVaju1b8DyXwy1yfE1jvMu3zeck4r1qmCKMHIRQfUCsPxX4gn0C3t5LeKOUyuVO/PGB7VxYahHK8PJ1JXV77fI7cTXlmdeKpxs7W3+Zn6l4DXUdSnuzeshmbdt2ZxXT2Fr9h0+C1Db/KQJux1xWL4U8RXGvrcm4hji8krjZnnOfX6V0VbYKlhWniMOvi9e5jjKmJTWHrv4fTsMm/1EmP7p/lXj9ja3o1S3LQTgeaucofWvY6MVOOy9YyUJOVuX/gFYHHvCRnFRvzAOlFFFemeaR3ECXNtJBMMpIpVh6giuE0P4T6fo/iGPUjeSzpA/mQxMoGD2ye+K7+mySJFG0krBEQFmYnAA9a1hVqQTjF7mVSjTqNSmth1ef8AxH8B33im5tr3S5Y/OhTy2ikbAIznINdRpPivRdcupLfS7+OeaMZKDIJHqM9a2KcJ1KE+ZaMmpCniKfK9V5HPeCPDcnhbw0lhcTCWYuZJCv3QTjgflXQ1l6z4k0nQBH/a14luZfuA5JP4Cr1pdwX9pHc2cqzQyruR1OQRU1OeT9pJblU+SC9nF7HhGsT+LB8QZtpvRdC6/cqm7Ztz8uB0xive4t/kp5n39o3fWl2jduwM+uKWta9f2yiuW1jLD4f2Lk+a9wooormOoKKKKACiiigArL8R6jNpWhzXdttMiEY3DI5NalVtQsINTsntboExPjIBxWVaM5UpKm7O2nqa0ZQjVi5q6vqcn4Y8W6hq+tJaXQiEbIxO1cHgV2joJI2RujAg1j6b4W0zSrwXVmjrKAQCXz1rZrlwNLEU6PLiZXlf8Dpx1WhUq82HjZW/E5qHwJpUFykyNPuRw4y/cHNeY+NfDfia68eXEsFrcTCaUG2lT7qr257YrsNc8YatY63dW1vJGI4nKqCgPFdto1zJe6LaXM5BkliDMQMc08px2Go1p08NCzW+hObYDE1qMKmJndPbUlsEni022S7bfOsSiRvVscn86sZHrVe/ydNucdfKbGPpXk+jRXI12zLRygecucqfWssdmDwtSEOS/N+H4G2CwCxNOcua3L+P4nsFFBOOtZ1r4h0i+vTZ2eo281wucxo4J4616ii3sjzHJLRsb4j1j+wfD13qXl+aYI9yp6noK4jwJ8Sr3xH4g/szU7aJfNRmieEEYIGSDntivRbq1gvbWS2uo1lhlUq6MOCK56w8N+GvBEdxqcUa2q7fnnlYttHoK6aUqXs5RlG8nsctWNb2sZRlaK3OP8ffDnV9b8UPqWltHLHchQ6yPt8sgAflxmvRdA01tH8P2WnySea9vEEZ/U0mj+IdK19JG0m8juRGcOF6r+BrSpVa1SUFSn0Ko0aUZurD7Rm+ItLfWvDt7p0cnlPcRFFf0Pv7V578P/h3q+heJv7S1Ro4o4VZUWN9xkJ4/KvVKKmGInTpumtmOph4VKkakt0FFFFYHQFFFFABRRRQAUUUUAFFFFAHL+L/ABFeaHJbLZiMiUEtvXPSp/CGuXWuWlxJeBA0bhV2DHar2raBY600ZvkZjGCF2tipNJ0Wz0WKSOxVlWRtzbmzzXlxo4v666jl+77X8v8AM9OVbC/U1TUf3nf5/wCRz2v6OtpfXFzIkkmlXxU3iRfeidekgHcY4NT2Phuwt9Pubvw9ctJNNbukMnmblBI4P510V9I0On3EqfeSNmGfUCvJbK7ubzX7b7PO2nvcTKkjWf7vdk4yR0J/Cs8XWwuGxEJVI3k9nYWFoYutQnGjL3Furmb4I8O+I7LxvDPcW89rFA5a6ll4Ur3575r1HwSMaFKygi3e7ma3z/cLnGPanN4V+0jZqerX97b/AMUEkgVX9jgDI9qq+OoSnhuKK1jIVZVCrGvQYPYV62YY2Tg67j8K27nk5dgV7SNFPd7nVZorznwIXttVuZbvdFEsBLPJkAcjua3brxhLJDJPpNgJLKL71/dy+TD/AMB7tXHgsX9apKpy2v03O/HYdYOq6blc6muO+IXjO48JWNr9igSWe5cgGTO1QP607T/GzzBZ7uG1lsTII3urKfzFhY9N6kAge9bmueH9N8SWK22qQCaMHcjA4Kn1Br06UoRqL2i0W6PNqOVSm/ZPUw/Afiybxlol2byFYZ4X8t2izggjgj3ritO+E+tW3iyKaWeH7HDOJPPD/MwBz09a9ASXw18P9NjtWljsYpWJAYktIe5Pc1uWN9balZx3djMs8EgyroeDXT7eVJylSVoyOf2EaqjGs7yiWK4z4keEr3xVpVsumugmtpC3lu2A4Ix+ddnRXLTqSpzU47o66lONWDhLZnGfDfwle+FdLuV1J0M1zIG8tGyEAGOvrXZ0UUVKkqk3OW7CnTjSgoR2QUUUVmaBRRRQAUUUUAFFFFABXF+KPFmoaPrTWtqIjGEVvmXJ5rtKx9T8L6bq14bm8R2kKhcq+OBXDjqeIq0uXDytK53YGpQp1ebERvGw/wAN6lNq2hxXdyFEjlgdowODTta0K112OJLwuBExZdjY61Z07T4NLsltbQERKSQCc9aq+I76bTdBnurUgSpjBIz3pyio4S2J96y187Exk5Yq+H92708rhovh+00ISizMh87G7e2en/661K4Twv4q1PVNdjtbt4zEysSAmOgru6nAV6FajegrRWhWOo16Na1d3k9QzRmuG+IqTPJp/kq7cSZ2gn+7Vr4eLKum3fnK6nzRjcCO1ZLMG8a8LyfO/lfsavAJYNYrm+Xzt3Ovooor1DzDzz4gfEO98MatDYadbROxj8x3lBPXoBWzp963j34dykD7LLeQvC2Oiv0yPar+v+ENG8StE+q22+SLhZEba2PTPpUU2veGvCSwaTLdwWQRBshGflHqcV2c0JU4qnH30cXLUjUlKrL3GcT4F+G+saJ4oTUdTeKOK3DbRG+4yEjH5V6tTY5EljWSJgyMAVYHginVjWrTrS5pm1ChChHlgebfEjwFqniTVrfUNKaNysQieKR9uMEnI/Ouq8FaBN4a8L2+nXUokmUs7legJOcCt+iqliJypqk9kKOHpwquqt2FFFFc50BRRRQAUUUUAFFFFABRRRQAUUUUAZF14W0e8uXuLi0DyyHLNuPJqvrmsJ4V021W2tRJGT5apuxtAFb9ZOvaBDr9vFFPNJEIm3ApjnjHeuGvQcKU5YaKU3107ndQrqdSMcTJuC6alXw14mPiB7hWthD5IB4bOc1v4Uc4ArG0Dw1BoDTNbzyS+cADvA4x9K070FtPuAoJJiYAD6VWF9vHDr6xrPW/9InFewlXfsNI6f1qRX6pe6dcWsdyqPNGyKwYZBIxXk3hP4ba/pvjK2ur1UhtrSXeZlkz5gHYDrz71NpVhfrrFqz21wFEykkocDmvX6eVZrVr05rk5en9bEZplVKjUg+fm6hXO+OPD03ifwxLYWsoim3rIm7oxHY10VFdsJuElJbo45wU4uMtmee/DbwLqHhi6ur3VJEWSZBGsUbZGM5yTXoVFFXVqyqz55bk0aUaMOSOwUUUVkahRRRQAUUUUAFFFFABRRRQAUUUUAFFFFADZY1mieOQZRwVYeoNZEHhPRra4jnhswskbBlO48EdK2aKynRpVGnOKbXdGsK1SmmoSav2OU17xm2i6s9mLMShVU7t+Ooz6Vt6JqX9saRDetEI/Mz8mc4wcVmaz4NtdZ1Jrya6mjdlA2oBjgY7itbSdNj0jTIrKJ2kSPOGbqcnNcGHjjVipuq/3ettu+nnsd1eWD+rQVJfvNL79tfxMO+VfEXiCTT5W26Vp5U3IzgTynkIfYDkimeO9Al8SeFjYaRLGssbq6xBgAwHb2rHvY5G0PWbRFZ7xdUaSRFBLFW5U/TFN8ERT2OrTT36SW8KwHMkwKqOR3NH9ozoY2FKML36mUMuhiMFOtOffQyfDPhHUfDHhnXrzXAIRPaGJLcMGyexPvnAFep6arppVosv3xCgbPrgVzzSt4w1KEQBholrIJGlIx9rkB+UL/sg85711VevUqyrVHUkeTQpRpR5Y7HnnxJ8C6j4nvLW90qSNnijMbRSNjjOcg10Xgjw7N4Y8MxWFzMJZtxkcr91SewroaK0liJypKk9kEcPTjVdVbsKKKKwOgKKKKACiiigAooooAKKKKACiiigAooooAKgvLOC/tXt7tPMif7y561PRSlFSVnsOMnF3W5l2PhvStOuluLO1EcqggNuJ61z+oeP3stRntRYB/JkKbvMxnH4V2lcpe+AbO9vprp7y4VpXLkALgZ/CvKxlHEQpqOBSjrrsj1MHWw86jljW5aabnS2kwu7OGcqB5iBsdcZFSM8cX32VM+pxTLaAW1rFApJEahQT3xXIfEO3uJ47D7NFJJtL7vLUnH3fSurE15YbDura7VtDmw9GOIxCpXsnfU7JJEk+46tj0OadXG/D63uIIL37TFJGSy48xSM8H1rsqrCV3iKMarVr9CMVQWHrSpJ3t1CvKfG3wz1XWvFUmo6bNE8V0VLiVsGMgAfiOK9Wor0KNadGXNA8+tQhXjyzKOi6edK0S0sGk8028QQuf4sVeoorJtt3ZskkrIKKKKQwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKADFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAZWpeH7TUbpbsNLa3artFxbvsYj0PqPrVUeD7OaRW1O5u9RVTlY7mYlPxUcH8a36KnlTJ5UNRFjjVI1CqowFUYAFOooqigooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAKGra7pOg26T65qllpsMjbEkvLhIVZsZwCxAJwDxWZF8QvBc8yRQ+L9BkkkYKiJqcJLE8AAbuTWnq2haTr1ukGuaXZalDG29I7y3SZVbGMgMCAcE815b4C8OeDrKPxzqmr6Do6waT4juyk8tjG32WGOKJgE+XKheSAO/SgD2GivLj8WfETWB1yD4a6tJ4a2+aL43cSzmLr5gtvvYxz16d61vEfxRtNFXwxNpumz6zb+JFkNqbVsSEiMPGAhHO4sFOSu3knoaAO7ory5fi9q9vqkmg6t4C1C18SyqJLDTIruOZLpOct5wwqBdpznOK2/DHxHXVINdj8U6RL4a1DQEEt/bTzrMqRMpZZFkUAMMK3QdvegDtqzxrmnHxI2gC4/wCJmtoL0wbG4hLlA27G37wIxnPtXBN8V9ZtYV1jVPAGpWPhZtrHVZLqIyxxsRiR7YfOq85POQO1SXuqWOkfHu/1LU7qO2srfwfHJLPI2FVftcnP+etAHoN9f2el2Ml5qd3BZ2sQzJPcSCNEGccscAcmrFeS/EjW7/xD+z74l1G+0d9JhmjU2kU8u6aSHzU2u6bR5ZPXbkkDqe1etUAV7O/s9RhabT7qC6iV2jZ4JA6hlOGXI7gjBHarFeap8SNP0L4Wz+Krfw/DZ2seqPby2kEgUZ+0mJ5cqnJPLkbck8Z70+L4parb+ItLtvEPgjUNG0jWLpbSx1Ga5jdmkf7iyRLzGW9Cc/rQB29prmnX2t6hpFrcb77TVia7i2MPLEoJTkjByFPQnGOcVoV5fZa9p/hr4pfEjVNXn8m2ht9KzgZZ2MUoCKOrMTwAOpNbGpfEDUdG8K6bf6l4XuF1fV7oW1jo8NyrSMzAsvmOQqxnapJHOPU84AO4orA8Ja5rWtWc58SeGZvD15C4XyHuUuEkUjhlkTg+47Vv0AFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV5ZpfhLUNZ8IfEvRZ4ZbGTV9bvTaSXEbIrq0UQRxxyhYYyMjg16nRQB5PH8QvFS+HxoL/DfXj4gW3+zh1iX7AW27d/n7sBe+Me2e9PtfB9/oGsfCjTvIlu49Gju0vbmKItHExtSMlsYUFzhc4zx3r1WigDib+wvH+OWj36Wk7WceiXMT3AjJjVzLGQpboCQCce1YeueE9Q8ReL/AIgWSwzW8Oq6Fa21tdyRsInkAl4DdDglcgdAa9SooA+bNK8FeEX+y6ZqfwU8SrrGViuJVuJ/se7gM4n87G3vwK7bxj8PrjxX8Wi0AurGOz8P272F/wCUzwJdRXTsqOD8smAQSp5wc+hr12igDyXxjqOv+L/gf4k02+8N6ha+IIESCa1itpHS4YSL88DAfvEIBPGSO/qd/Rfij/bOtW2nf8IN40sPtD7PtN9pHlQxe7tvOBXd0UAeJHQNXb4GHTzpN6bs+IvONsbZ/MMf9ob9+3Gdu35s9Mc9K7f4p2F5qHh/Sk0+1nuZYtcsZmWCMuyIswLMcdAByT2rtqKAPHNW8B6trXxX8U67pols9Q046ddaPPcxMbW5kWGVXRgflYcgbh8yE5BHfY1/xDH4i+H9vP4j+H2vXpFz5d5p0Fs32m0lUH97CflLjPR0PRvqK9LooA8++FR15o9XbUrfWrbRGljOkQ68wa9Rdp8wPyW2527QxJHNeg0UUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFVrzULTT0V72dIVY4UucZNTKUYq8nZFRi5O0Vdlmiqdnq9hfymOzuo5nUZKoc4FXKIzjNXi7oJQlB2krMKKKKokKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiqd5q9hYSiO9uo4XYZCucZFPs9QtNQRmsp0mVThihzg1mqtNy5FJX7GjpTUedxdu5ZooorQzCiiigAooooAKKKKACiiuD8cTCLVDI14mILQv9lku3tW6n542HDNxjBz29aAO8orhZvGWoxRXEEFuvnW0SXBadSMxOE8st/tZZgf901m694g1sSXFs1/DDJbNcQ5jQqJiI4nU9eCN/6UAemUVwN/4x1ewuWsZFt2mjlmDXO0KhCKjBcMwAJ3+vRTxVNfFup291cXEzxyPh3SEtxbgpEecHlRvJz7UAelUVm6Jeve2s3m3dvdtFMY/NtgQp4B9+eexrSoAKKKKACiiigAooooAKKKKACiqN1rWnWU5hu7yKKQDJVmwantL22v4fNs5lmjzjcpyM1mqtNy5VJX7GjpzUeZxdieiiitDMKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK5Lx/Z3F5YWi2sDzFZSSEXOOK62iufFUFiaMqTdrnRhq7w9aNVK9jgPAenXlnrM73VtLEphwC64BORXf0UVng8LHCUfZRdy8ZipYqr7WSsFFFFdhyBVWfU7G1lMVzdRROBkqzgGrVeZ+N7K5n8TSPDbySL5afMqEjpXn5hip4Wj7SEbu9jvwGFhiq3s5ysrXPR7e5gu4/MtpUlTONyHIqWua8CQyQeHik0bRt5rHDDFdLXRhqsq1GNSSs2jDE0lRrSpxd0mFFFFdBzhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB558Rbeb7ct55Tm3gti0koUkIASST+Fafw7gki0q4kkQqkrq8bEcOuOo9RXXsquhV1DKwwQRkEVgGOy8D6LNLbw3clgJt7RR/vBbK3UqvUIOuBnHOK8uGWwhi3iubV9PkenPMZzwiwvLoup0FFZV14n0WytLa6utTt44Lpd8EhfiQeoP41JpfiDStaeRdKvobpogC4jbO0HpXp3V7Hl3V7GjRWbrXiHS/D1sk+r3a26yHbGuCzSH0VQCSfYVWutQ1bUdDt7rwzbQxzXB/5iivF5Sc/MUAyT7cdaYy8+sadFIySXsCspwQXHBq3HKk0ayRMHRhkMDkGvF7+y1CLUHjvx511JIwMkcRVZWHJKjnivWPDyNH4eskdSrCIAgjBFePgcfVxNadOcLJf5nr43A08PRhUhK7f+RpUUUV7B5AVBdQ2ciq97HAwQ5VplB2n2zU9c/wCMNHudZ06COzj3vFLvwJVQj5SOjKynr0IoA0LewsrfULqfzfNuLrBcSuGwo6ADsvU1Pmxmcf8AHvI2SR90nOMH9K4KTwf4gmuA8wtFcLtaSJwikeTsx93dnd6tjHQVam8GzWiXFxbwwIytE6lGOQFgKyY46k/nQB1l1c6b5tvBcLDKbmbYg2BwXCk8+hwDU5+w7WlP2fDfKz/Lz2wTXn1l4S1XULC1ngjh0rNqsY2SEtu8qRfN6dcyLx1+WnHwVqp0x18jEhdSsYuY8DCFd2PL2EEnoRnHfNAHosUUcMYSCNY0HRUUAD8BT6r6fFNBplrFdFDPHCiyGMYUsAM49s1YoAKKKKACiiigAooooAKKKKAPN/GmmXt14jkkt7SWVNijcqEiuk8DWs9poDR3MTxP5zHa4wccV0lFeXRy2FLFSxKlq76ep6dbMZ1cNHDOOitr6BRVO81ew0+RY726jhdhkBzjIp1lqVnqIc2NxHOEOG2HOK9D2tNy5OZX7X1OD2VTl5+V279C1RRSN90/StDMpHW9MViDfQAg4I3irqsHQMhBVhkEdxXjFzpt4bqYi0mIMjY/dn1r1/TgV0u1VhgiFAQe3yivHy7H1MXOUZwtY9fMMBTwsIyhK9yzRRRXsHkBRRRQAUUUUAFFFFABRRRQAVxvxG8YXPhPS7Y6fGrXN05VXcZVABzx3PNdlWZr3h/T/Eenmz1SHzI87lYHDIfUGtaMoRqJzV0ZVozlTapuzOV+G3je+8Utd22qJH50ADrJGu0EHjBFd7WN4d8K6X4Xtni0uIqZDmSRzlm9Oa2adeUJVG6asicPGpGmlUd2FFFFYm4UUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAebfEX4g6l4c1iHTtJSJW8sSSSSpuznoAK6PwF4nm8U+HRd3UYSeOQxybR8rEdxU/iTwZo/ilon1OJvNiGFliba2PT6VpaRpFloenR2OmwiKCPoO5PqT3Ndc6lF0FGMfe7nHCnXVdylL3exdooorkOwKCMjBoqvfX9rplm91qFxHbwJjdJI2AM9KAPNfiv4furm20mDQ9Mkkig8zMdtFlUzg9B071j/DvwZq893fLfz6po9s0abhCvlGfk8biMjHt616hPpF7qHiKO8u79l062Aa3tICV8x8ctIe4HZenc03wvo91oVpd2M0qyWoupJLPklkiY7tp+hJH0rL2S5+cy9mufnNS2sYLW1ggRS626hY2kO5h+J5zViiitTUyNU1j7Br+i2HkLJ/aEkqeYTzHsjLZH1xitesnU72xt9e0a3u7fzbq5klFrJtz5REZLHPbI4rWoAKKpXWr2lrJ5Rcyz/8APGEb2/IdPxqDdq199xU0+I/xP88h/DoP1oAv3F1BaQmW6mSGMdWdgBVD+1Li7/5Bdmzp/wA97jMafUA8n8qkg0a0glFxMGuZ15864beV+meB+GKDu1RioJWzHBI4M3t/u/zoAzk0641iYPe3sktqjZIi/dpIR2AHJHuTzXQgYGB0pFUKoVQAAMAAdKWgAooooAKKKKACiiigAooooA8s8efEjVtD8SvpukpCiQqpdpE3FiRmu28G6+/iXwvbajNGI5WykgXpuBxke1QeIfAmieJbxLrUIXWdRgyRNtLD0Nbenada6Vp8VlYQrDbwjCIO1ddSpRdGMYR97qcdKnXjWlKcrx6FmiiiuQ7AooooA4Px5pt7e6rbvaWs0yrDgmNCQDmrngCwu7GG9F5byQFmXb5ikZ4NdhRXlRy2EcZ9b5nft8rHqSzGcsJ9V5Vbv87hRRRXqnlhRRRQAUUUUAFFFFABRRRQAUUUUAFeP/EH4ga7pniubTtLm+yQ24UfcBMhIBzz25xXsFYuteEdE8QTxz6rZLNLHwHBKkj0OOtdOGqU6c71FdHLiqdSpT5acrMj8Fa3ceIfCdpqF5GEncMr4GAxUkbh9cVvVHb28NpbR29tGsUUa7URRgKKkrCbTk3FWRvBOMUpO7CvHviB4/17TPFk2n6ZP9lgtwo+4CXJGc817DWLrPhDRNfuo7jVLFJpo+A+SCR6HHWt8NUp053qK6MMVTqVIWpysxPB+sz6/wCFLLUbuMRzSqQ4A4JBIyPrjNaN9qVppqK99OsKucKW7mpoIIrW3SC3jWOKNdqIowAK5fx7Y3V9YWi2dvJOyyksEXOBivOx1eVKlOrSjdrZfM9HA0Y1akKVWXq/kb1lrOn6jMYrK6jmdRuKqegq9XA+BNLvrHWJ5Ly0lgQw4DOuATkV31Y4DEVMRQVSrGz7G+OoU8PWdOnK6MfW/EtpoMkSXccrmUEr5YB6fjS6J4ktNeaVbSOVDEAW8wAdazPGHhy91y4tnsjEBEpDb2x1NL4P8O3uhyXLXpjIlChdjZ6Vye2xv172fL+7728u/qdPscH9S5+b952v59vQ6miiivZPICiiigAooooAKKKKACiiigAooooAKKKKAPCtX+KPiSPxHcG3mWG3hmKrbGMEYBxgnrzXt1lcG6sLe4ZdhmiVyp/hyM4rIuvBXh+91ganc6dG9zncW5wx9SOhreAAAA4A6V14irSqRiqcbWOPD0a1OUnUle+wUUUVyHYeEX/xT8SDxBLJDKsNukxC2hjB+UHGCeua9ytZjcWcMxXaZEVivpkZxWNceCvD91rP9qT6dG11u3FucE+pHQmt4DAwK68RVpVFFU42sceGo1qbk6kr3MzWPEOnaHBJJfz7WjiMojUFncDsB3PFULSLSfF7WOrypJI1qodLSV8rC55yyjguPXnFc98RbW4+2i+EEjW1vbFpZVUkIASST+Fanw7t5odKuJJY2VJnV42I4dSvUeor5ynja8sa8O4e7rrr2+4+jng6EcEq6n72mmnf7zr64LxlrYsfFGk3Nl532rT5iJ484Sa3kGHXr1BCsPdfeu9rzbx5pdymuW9yoV/7QuI7W3jVvmZyOTj0ABJPoK6MwqYmnSTwyvK/roc+Ap4epVaxDtG34na2PiLT7zSzfvKttACQWuGC4x+NJq/iOw0WOFrozSm4/wBSltC0rP8AQKDWPpHgXTvscP8AwkGn215cwsdhf51UZ9Dx+ldYqqqhVAUKMAAdBXThpVJUYuqvetqc+IjTjWkqfw30PnTXvGmt3fiT7Wl9cw/Y55GtEljCtCGBUggj045r1rwjrFv4v0G2WbUrg3UEYF3b58p2J/iPAO09iODUd94B8I6l4kZZlP27abmaBZj86sSAxHpkHFdLaaDp1gYJLO0jSa2hMEUpGWVP7uepGadOE4ybkzihGUZNtlq1srayj2WsKRL/ALI6/wCNTMwVSzEAAZJPasbRNZu7p7m01qwaxvbTl2GTBKh6Oj+nqDyKtANqjBmBWzByqngze5/2f51sbB82qPzlbJT06GY//E/z+nXQACqAowB0A7UAYGBwKKACiiigAooooAKKKKACmTyeTbySBSxRS20d8DpT6KAPCIfil4kbxAkjTKYGnANr5YxtzjbnrXuyNvjVsYyM4PasJfBPh5dZ/tRdNiF1u35527vXb0zW9XXiatKpb2cbHHhaVWnze0le4VFdTG3tJpgpcxoWCjvgZxUtHXrXIdh4Pa/FLxK2vxyyTK8DTAG18sY2k4xnrXuyNvjVsY3AHB7VhxeCfD0Os/2pHpsYut27PO0H129K3q68TVpVLezjY48LSq07+0lcoXeuabYXBhu7yOKQDJVjzU9nfW2oQedZTLNHnG5fWuB8ZaRqF54ieW1sppoyijciZFdH4HtLiy0ForuF4X85jtcYOOK+coY2vUxkqEoWir669D6Ovg6FPCRrRneTtpodHXN6h430/TtQmtJoZ2eJtrFVGP510lcBrfgvVNQ1u6u4DB5cr7l3Pg9PpWuY1MVTpp4ZXd+19DLL6eGqVGsS7K3e2p2umahFqunRXkCsscoOAw54JH9Ksn7p+lZ3h6wm0zQbazudvmxht205HLE/1rSPNdtFzlSi5/FZX9TirKEaslDa7t6Hk1x4s1xLmVV1CQBXYAbV9fpXqVhI0unW0kh3O8Ssx9SQKyH8GaG7szWhJYkn94a3Io1hhSKMYRFCqPQCvNy7CYrDzk68+ZPbVv8AM9LMMVhq8IqhCzW+iX5DqKKK9g8gKKKKACiiigAooooAKKKKACiiqn9raeLz7Ib23+0Zx5XmDdn0xTSb2E2luW6KKKQwooooAKKKKACiiigAooooAKKKKAKWsXElro13PAdskcRZTjoa87sfGOtTX9vHJdAq8iqw2DoTXpl1bR3lrJbzgmORSrAHHFYcXgjRoJkljilDIwZf3h6ivGx+GxdarGVCdkt9Wj2MBicJSpSjXjdvbRM6GiiivZPHCiiigAooooAKKKKACiigkKpLHAHJJ7UAFFVLbVbC9maK0vYJpF6pHICR+FW6bTW4k09gooopDEIDKVYAgjBB71i6ldalpGoWZsrFbnSWxDLHAuJICTgOB0K+o7da26KADOOtYel60mta/qdvFbo1tpkiwpck5LTFSXA9NoKjPqTXG/FSy1bT9LuL6x1F20+6lTz7aVzmFxwGibsD3Xp3FcZ8LJpf+E/s0Er7GErMu44J2nkj1rGVXlmo2MpVOWSjY+gaKKa52xsQQpxwW6CtjUytO0u3XxFqetRXIuHvEit8LgiIRbsrn/eZia16x/CuiN4f8OwWE0qzzhnlmlUYDyO5dj+bVxnxA+IOseGPES2OmpatEYVcmaMscnPoRUTmoK7JlJRV2drfWUuuWs9vLLNaW7AqpjO2Qn+9z2B7d657Vdd1Xw/oC2mrX9p/avnbLeSMgNcxD+Mp2Pr2rP8AB3jTxP4qiKR2NpGEmAlvXBWNU6lVXJLOfwArsb/wzpOpTSz3NmhuJgoknUYdgvQbuuPasa6nVoSVJ2bWh0YWpTVWM5q8b6nLeFfE2qal4igtbu4DxMrkrtA6KTXf1i6d4U0vS75Lu0jkWVAQC0hI5GK2q58uo4ijRccRK7v3vpodWYVqFaqpUI2Vu1tQooor0TzwooooAKKKKACiiigAoopruscbPIwRFGSzHAAoAdRVW01Oxv2ZbK7huCv3hHIGx+VWqbTW4k09UFFFFIYUUUUABOOtGQehrkviHI8ej2xjdkPn9VOP4TWD4Cnlk8SkSSuw8huGYnuK8irmSp4xYXl3trfuetSy11MI8TzbX0t2PS6KKK9c8kKKK4fxH4w1LSdcms7VbcxoFI3oSeRn1rkxWLp4SHPU22OrC4Wpip8lPfc7iub8aateaTp9vLYyeW7y7WOM8Yq54X1W41jRVu7sIJDIy/IMDAqzqujWmswpFfKzIjbl2tjmorOeKwjdB2clp0LpKOGxSVdXUXr1OX8G+IdS1XV5Ib6cSRrEWA2gc5rt6ydL8Nado9y09kjq7LtJZyeK1qWApV6NHlru8vW48dVo1a3NQVo/cFFFFd5whRRRQAUUUUANlVmhdY22uVIU+hr53bwl4m/4S4wGyuTd/aN3n4O3rnfu6Yr6Korqw+JlQvZXucmJwscRa7tYbEGWFFkO5goDH1NOoorlOsKKKKACiuf8X6xdaNpkU9kVDtKFO4Z4wayvCfijUdX1k2140Zj8st8q45FefUzCjTxCwzvzP9Tvp4CtUw7xCtyr9DtaKKK9A4DG13xLbaA8K3MUknmgkbMcYo0LxJba+0y20UkfkgE78c5qh4v8O3muyWzWbRgRAht5x1o8IeHbzQpLprxoyJQoXYc9K8f2uN+vclv3fe3l39T1/ZYP6jz3/edr+fb0Ooooor2DyArLbxNoyMVbUYQQcEZ6VqV49c6DqrXUxXT7gguxB2e9eTmWMrYVRdKHNe/f9D1cuwdHFOSqy5bW7fqevxussayRkMjAFSO4p1VtORo9LtUcFWWFAQexwKs16kW3FNnmSSUmkFFFFUSFFFFABRRRQAVk+KLS8vvC9/baacXMkJWPBxk+ma1qKqMuVpkyjzRcX1PBvh/4a1+LxtazCzuLSO2fNxJIpUbe689c+le80UVviMQ68uZqxhhsOsPDlTuFFFFcx0hRRXIeNPEN/o9zBDYuqLNExYlckHOOK5sTiYYak6tTZHRhsPPE1FShuzV1R4NettQ0fTry3F7CFEpaPzRAW6EjpuxnA+lUoPB0ehaCtt4T+zWd/G29bq5h80yNn5t3Q/N6jp2rI+HGqWlvHJozSQxztmaNCf3k5/jck/ebJGa76nh68MTSVWGzJxGHlh6rpz3Riw63NpuhfbfGC2ulOknlyOk2+I5OAwYgYB9+lc3498RxXWkz6NY7ma4MW+eNxt8osC20jvgEfjXeSxRzxNFMiyRuMMjjIYehFeW6h4Vj/wCE0m0bw9bRQRLbi6ZVJCRZOMexJyQB6Vz5hUxMKSeGV5X/AAOrAQw86rWJdo2/E7aw119c0S5k8PxBLqH93GL1GWPdgEcgfMMHt9K5jVvhbJrkct7rviGebU2T/WrEiwxY7KmM7R7nNdb4Y0ufR9FW0uipkDs3yHI5q7qtkNS0e8sicC5geLP+8pH9a3oc9SjF1l71lf1OTEwpe1lGGsb6HN2fw9sLHw3Lpun310jNKLiC63DfDKB8rDAAI9j1Fb8F+lpbx2+rXtt9vjtxLOIztyOhcKeQuaXQbS4sPDun2d6Va4t7aOKQocgsqgEj8qy/GXh2LWtN+0ReZHf2oLQSw/fIPVD6qe4rWb9nTbitkRShFyjHoaEfiTR5ZFSPUIWZiAoB6mtOvI9O0LVI9TtXewnVVmUklOgzXrledluMrYqMnVjy29f1PSzHCUcLKKpS5r+n6BRRRXqnlhRRRQAUUUUAFFFFABXOePtPv9U8F31rpIZrhgp2KcF1BBKj6iujoq4ScJKS6ETipxcX1PEvhboGtW/i5buS0ntbWFGWZpUKBsjheevP8q9toorXEV3XnztWMcPh1h4cidwooornOkKKK4rxZ4o1HR9ZFtZtGIzGG+Zc81y4rFU8LT9pU2OnC4Wpiqns6e52M1vDcKFniSRQcgOucU2KytoH3wW8cbYxlUANY3hDV7rWdLknvSpdZSo2jHGK36qjUp14KtFbk1oVKE3Sk9grmL/x1Y6fqE1pLbTs8LlSVxg109cBrPgnU7/Wrq6heERzSFl3Mc4rkzKpiqdNPCq7vrpfQ68up4WpUaxLsreh2el6jHqunR3kKsiSZwG6isnVfBljq+oyXlxPOjuACEIxwMelX/D2nzaXocFpcFTJHnO05HWtOt1RjiaEViI3ejfqYe2lh68nh5WWqXoUdH0mHRdPFpbO7oGLZfGefpVm5uobO3ae6kWKJfvO3QVLWN4st5brwzdRW8bSSMFwqjJPzCtKn+z0H7NfCtF6EQ/f117R/E9X6liDxBpV1OkNvfQySOcKoPJNaNeW+HdG1K38RWUs1jMkayAszJwK9Srky7FVcVTcqseVp+f6nVmGFpYaoo0pcya/rYKKKK9M80KKKKACkZgilnOFUZJPYUtR3EK3NtLBJnbIhQ49CMUAcdbfFTw/c60uno0y7pPLWdl+QnOPyrta8es/g3qEOvxvPfW5sI5Q+9c72UHIGMYB/GvYAAqgDoBgV14mNCLXsXc48LOvJP2ysLQSACTwBRTZEEkbI33WBBrkOw43/hafh7+2/wCzt8uPM8vz9vyZ6fl712gIIBByD0NePH4NX51/P26D+z/N37+fM25zjbjGfxr1+OMRRJGv3UUKPwrrxMaEeX2Tv3OPDTry5vbK3Yo6zottrlqkF2zqiPvGw4Oap6R4TsNFvftVq8xfaVw7AjBrcorzJYWjKoqsormXU9SOJrRpukpPl7BRRRXQc4UUUUAFFBOBk1nnX9KBIOoW+Rx98VEqkIfE0i405z+FXNCikVgyhlOQRkH1pasgKKKKACiiigAooooAKKKKAK2o6jbaVp817fSCOCFdzMa5zw78RtF8R6n9gtfNhnIJQSrjfj0961vE2hr4j8PXOmPKYjKPlfGdrDkcVwngz4W3+ieI4tS1W5gKW2TGkJJLnpzkDArrpQoOlJzfvdDjrTxCrRUF7vU9RooorkOwgvbyDT7KW7u5BHDCpZ2PYVzGg/EnRPEGrDT7bzopnz5ZlXAfHYVu6/pCa7oN3psrmNbiPbvA+6eoP51554R+FWoaP4mg1DU7u3MNq++NYSSXPbOQMCuujGg6cnUevQ46066qxVNXj1PVKx9a8NWeuzRS3jSq0SlV2NjjNbFFefVpQrR5Kiuj0KVWdKXPTdmc3H4F0dIPLdJXKypNHKXw8Tr0KsOR1wfUVeh8Q27eI5tFuo5LW6C+Zb+bjbcpjlkPfHcdRWtUclvDNJHJLEjyQkmNmXJQkYJHpxTp04UoKEFZIKlSdWTnN3bCSeKFGeWREVASxZsYA71leFNal8ReH4tWlgWCO6d2t1ByWh3EIx9yAD+NeD+NZtSXxhqcd9NmTzSGEJYIR24PtjNejfDHV9YtvD1jZy6DLLpzyMkV9b3CtsGf44zgqB6jNTCrzScbHLGpzScbHpdFUDrulDVX0xtQt1vkwTbtIA/IyOD14q5LLHBC0szhEUZZmPAFatpK7N0m3ZGP4Qg1G18PJBrBkNzHPMN0r7mZPMYoc/7pFbdct4d8SpdaxrFle30cnlXoFoeBujZAQB64ORXU1MZxn8LuOUJQ+JWCiiirJCiiigAooooAKKKKACiiigCjrOsWmhaXLqGoPshiHOBkk9gPesfwx490jxTdSW1kZIp0G4RyjBYeoqz4x8Of8JT4cl05ZhDIWDxuRkBh0z7VyngH4bXnhvW21TVbiFpERkijgJI54JJIHbtXXTjQdGTk/e6HHUnXVeMYr3ep6RRRRXIdhT1XVLXRtMmv7+Ty4IRlj6+w96wvDXxC0fxPftZ2fmxThdypKuN474rQ8WeHx4m8OXGmmXyWfDI+MgMDkZ9q4vwN8Mr7QPEC6pq1zA3kKRFHASdxIxkkgdq66caDoyc373Q46s66rRUF7vU9NrD1fwpY61ei6unmDhQvyMAMCtyivPq0adaPJUV0ejSrVKMuam7Mz9G0a30S1a3tGdkZtx3nJzWhRRVU6cacVCCskTOcqknObu2JvUdWH50oOelePa9NMPEWoASSAC4fADH1r03wyS3hjTyxJJhGSa8vB5l9arTpctuXz87HpYzLvqtGNXmvzeXlc1KKKjuLiG1hMtzIsUa9WY4Ar1m0ldnlpNuyJKKpQazp1zMsUF7DJI3Cqr5Jq7UxnGavF3HKEoO0lYKKKKskKKKKACiiigAooooAKKKKACoFvrVrk263MJnHWMSDcPw60t4sz2Fwlq2ydomEbejY4P51876ZoHiNfF0Kpa3Ud8twGeZlIwc8ktXXh8PGspNytY48TiJUXFKN7n0dRSLkKM8nHNLXIdh5x4m+LC6H4gm06108XK27bZJGk289wK7vSdTg1nSLbULXPlXEYdQeo9q43xF8KrHXdefUUvJLbzm3TRqoO4+o9K7XTdPt9K02Cxs12QW6BEGewrrrOh7OPs9+px0ViPaS9p8PQnlkWKJ5H+6gLH6CsBfHGiu4UTSZY4H7s1t3iNLZTRoMs0bAD3xXmEXg3W1uEY2nAcE/OPWvnMxxOLoSisPG6e+jZ9Hl+GwtaMnXla22qR6qDkZFFIowoHtS17B5A2TmNsehrxqXSNQNw5FjcEbz/wAsz617PRXl5hl0cdy80rWv+J6eAzCWC5uWN7kVoCtnCCMERqCD9Kloor00rKx5rd3cKKKKYgooooAKKKKACiiigAooooAKCcDJ4FFYPjW21C88H38GkFvtTR/KEOCwzyB9RmqhHmklexM5csXK17GvBe2t0zLa3MMxX7wjcNj8qnrwr4YaPrMPjSGdLae3t4wwuGdCoIx056817rW+JoqjPlTuc+FruvDmasI7iNGdjhVGSfavL0+M8Da2IP7OxYmXZ5+/5tucbsfrXqDKHUqwyGGCPWvO1+Dulrr32v7XKbMPvFrj3zjPpVYZ0Pe9t8icSsQ+X2PzPRQcgEdDVbUNQg0yye6uyViTGSBnqcVZAwAB0FY/imxn1Hw9PbWib5WK4XOM4NediJThSlKCu0nY9KhGE6sYzdk2rnLabqWn638Qrua4hU2xhSC0jeFT5rH5ndvfoBmu+hgit4xHbxpEg6KigAfgK888PeFtWsdftLm5ttkUb5Ztw44r0auLLq+Ir03LER5XftY68woYehUSw8rq3e5RvtF0zU5I5NQsLe4kjYMjyRgspHQg9RWL4u0XULy3mu9P1a6gVYtslkqK8cwzyMEZBPqK6iivQqQ54OHdWOGnPkmpdmeR6Lpd/Hrlk8llOqrMpJMZAHNeuUUVw4DARwUXGMr3O3HY6WNkpSVrBRRRXonnhRRRQAUUUUAFFFFABRRRQAUUUUAFIzKilnIVQMkk4Apa5P4k2epX3gy4i0kOz7gZEj+86dxV04qc1Fu1zOpJwg5JXsdNb3dtdqTa3EUwU4JjcNj8qmrxn4Q6Vq9t4kmuHt5rey8krL5ilQ7dhg9TXs1a4iiqNTkTuZ4as61PnasQ3l3FYWM13cHbFChdz7AZrzfSvjFHfa7FaXOneRbTSbFlEmWXPQkV6TdW0V5aS21wu+KVCjr6gjBrgdL+EOnafry30l5LPBE++KBlxg9snvV4d4fll7XfoZ4hYjmj7Hbqeh1T1PVLbSLT7Tesyx7guVXPJq5WD4w0651TQvs9lH5knmq2M44Ga8zEznToylTV5JaHp4aEJ1oxqOyb1G2GuaDrF+Le2iWSdwW+eAc+vNb6IsaBI1CqOAAMAVwXhTw3qmm+II7m8t9kSowLbgeorvq5cuqVqtJzrx5ZX7WOrMKdKlVUKMuaNu9wrD8YQyXHhi5jgjaRyVwqjJPNblFdtamqtKVN9VY46NR0qkai6O55Z4W029h8TWck1nNGiucsyEAcGvU6KK5MBgY4Km4Rd7u51Y7GyxlRTkrWVgooor0DgCiiigAooooAKKKKACgnHWivGfi9qGsQeI4YhLNBYiEGExkqrN/FkjvW+HouvPkTsc+IrqhT52rns1GOc1yPwyudRuvBUD6qZGfewiaT7zJ2NddWdSHJNx7GlOftIKfcKKKKg0Ciiqup3y6bpdzeupZbeJpCo74FNJt2Qm0ldlqivHtB+Leq3fiS2gv4ITaXEoj2xrhk3HAOe+K9hratQnQaU+phQxFOum4dAqC+vI9P0+e7mz5cEZdsegGanqOeCO6t5IJ1DxyKVZT3BrFWvqbu9tDy7R/jFNf+Ibe1utOiitLiURh0cl0ycAnsa9VrhdK+E+jaXrkeorNPMsL+ZFA5G1SDke5xXdV04mVByXsUcuFjXUX7Z6hRRRXKdZ5Prfxgu7HxDPa2WnwyWlvKY2MhO98HBx2Feoafex6jptvewgiO4iWRQeoBGa5DVPhVoeq60+oSPcRea++SGNgFc9+2Rn2rs4II7a3jggUJHGoRFHYDgV115UHGPslr1OPDxxCnL2ruuhJRRRXIdgUUUUAFFFFABRRRQAZA70V89+ONU1xPG98s1xcwskuIEViBs/hwK9y8Oy3k3huwk1NSt20CmUMMHOO9ddbDOlCM73ucdDFKtUlBK1jSxjpRRRXIdgUUUUAFFYHjTxE3hjwzNqEUQll3CONT03HufauS+H/xH1DxD4gOmatFEfNjZ4niXbgryQfbFdEMPUnTdRbI554mnCoqT3Z6ZRRRXOdBj+KvECeGfDs+pPF5rR4VI843MTgc1yXgn4mzeJdd/sy/sooGkQtE8TE8jqDn2rttb0a11/SJtOv1JhlHO04II6EVz/hX4c6Z4W1Fr6Gea5uNpVGlwAgPXAFddOVBUZKa97ocdSOIdaLg/d6nX0UUVyHYRXVwlpZzXMudkKM7Y9AM15VpnxluLnX4obvT4ksJpAgKEmRQTgH0NesSRrNE8cg3I6lWB7g1xVh8KdBsNcXUUaeQRvvS3dgUU9R2ya68PKgoy9qrvoceIjiHKPsnZdTt6KKK5DsCiiigAooooAKKKKACiig9OOtABkZxnmivnPXdW8RL4xufNnu471bgiOJSR3+UAenSvoazaZrGBrobZjGpkHo2Of1rrxGGdGMW3e5x4fFKvKSStYmAx0ooorkOwKKKKACiuW8f+K5fCegpcWkSyXM8nlx7x8q8ZJNYvw6+IF74m1C4sNUij81Y/MjkiXAxnBBFdEcPUlSdVbHPLE041VSe56HRRRXOdBzvjbxUPCWhfbVgE80kgjiQnAJ65J+grF8B/ESTxXqE9je2kdvOieYjRMSGGcHr9a6bxJ4csvE+ktYahuC7gyOhwyMO4rM8JeAtO8JSyz20stxcSrtMkuOF9ABXXGVD2DUl7xxzjiPbpxfuHU0UUVyHYU9Y1KPR9GutQnUsltEZCq9TjtXnPhj4tXWreIoLDUbGGKG5fZG0JO5Cemc9a9MvLSG/sprS6QSQzIUdT3BrkdD+F+i6FrK6jFJPO8ZzEkpBVD68Dn8a66MqCpyVRa9DjrxxDqRdN6dTtKKKK5DsCiiigAooooAKiuLW3u0CXUEcyg5CyIGAP41LRRsG4iqFUKoCgDAAHSloooAK4f4ranqmm+Fo20ppIlkmCzzR9UXHHPbJruKZLFHPE0U8ayRuMMjjII9CK0pTUJqTV7GVWDqU3FO1zyv4P6zq9/f31veTzXNosYYPKxba+egJ9q9UliSeF4plDxupVlPQg1HaWNrYQ+VY20NtHnOyFAo/IVPV16qq1HOKsRh6TpU1CTuclpfw18P6TrQ1K3hkaRW3RpI+UjPqBXW0UVnOpOo7ydzWFOFNWgrBRSM6p95gv1NLUFhRRWX4ljvZfDOoR6USLxoGEW0859vfFOKu0iZOybL8d3bzSGOGeKR16qrgkfhUtfPngHT9ZXxzZGCC4iMUn79mUgBe4Oa+g66cTQVCSinc5sLiHXg5NWCiiiuU6wooooAKKKKACiiigAooooAhls7WeZJZ7aGSRPuu8YJX6E9KmooouFgooooA8p+MGs6vYXtjBZzzW1o0ZYvExXe+ehI9B/Ot34U6nqmpeF5G1VpJVjl2wSydXXHPPfBrsruytb+Hyr62huY852TIGGfoakihjt4VigjWONBhURcAD2Fdcq8XQVLl17nHHDyWIdXm07FfVNLtNZ06Wx1CISwSjDKf5j3rG8N+BdG8L3Elxp8cjzuNvmStuKj0FdJRXOqk1FxT0Z0OnCUlNrVBRRSF1DBSwDHoM1BoLRRRQAjMqKWchVAySTgCmQ3MFypNvNHKAcEowbH5VyfxPt9SufBcyaUJGPmKZkj6tH3/AKVxPwes9Vj8RXE5jmjsvJKyl1IDN2HPeuuGHUqDq823Q454lwrqly79T2aiiiuQ7AooooAKKKKACiiigAooooAKKKKAIWs7Z7hbh7eJpl+7IUBYfj1qaiigLBRRRQB418Wdc1mz8TRW0FzPa2iwho/KYqHJ6nI613Pw11DUdS8F282ql3kDsqSP950B4J9frXSXmnWWoKq39pBcqpyomjDgH8amRFjjVI1CIowFUYAFddSvGVFU1HVdTjp4eUK8qrlo+hQ1zQ7HxDpj2OpxeZExyCDgqfUHsao+GvBuk+FVl/s2N2ll4eWVtzEen0rfornVSajyJ6HQ6cHPna1Ciik3ru27hu9M81BoLRRRQA2SRIkLyuqIOrMcAUkM8Vwm+CVJV6bkYEfpXB/F221O58Mw/wBnLI8CTZuEjGSRjgn2zWR8GLTU4JNRluI5Y7F1UKJAQGfPUZ9q61h06Dq83yOOWJaxCo8vzPVqKKK5DsCiiigAooooAKKKKACiiigAooooAKKKKAPL/FPxYudH8STafp9lFLFbNskaUnLHvjHSvQtF1SLW9FtdSt1ZY7mMOFbqPauZ1/4YaPr+stqMss8EkhBlWMjDn19q62xsoNOsIbOzQRwQIERR2ArrrSoOnFU1r1OOjHEKpJ1H7vQnooorkOw8P+LNzqyeLtjvMloIl+z7CQp9enfNelfD2TUJfBFi2rb/ADsMFMn3imflz+FdHLBFNjzokkx03qDinjjpXXUxCnRjT5djjpYZ060qvNe/QKKKK5DsEAA6ACloooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAGyOI42dvuqCTXkyfGeZtdCmwjGnGXbnJ8wLn73pn2r1ogMpDDIIwRXDJ8JdCTXPt++YxiTzBbZGzOc49ce1deGlQXN7Zehx4mOIly+xdu53QOVBHQ0UDjpRXIdgV85eJL3xAPGl0bmS6W9W4IjVCeBn5do9MYr6NqNreF5RI8SNIvRyoJH411YbEKg23G9zkxWHeISSlaxDpbXLaTatfjFyYVMo/2sc1aoormbu7nUlZWCgADoMUUUhhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAV7+8j07T57yfPlwRmRto5IAzXl+kfGK4u/EEVveWEUdnNJsBQkumTgH3r1SeCO5t5IJ0DxyKVdT3B61xWm/CjRNN1tNQWSeVY33xwORtU9vriuvDyoKMvarXoceIjiHKPsnp1O5ooorkOwa+7y22fewcfWvnB73xEPGJbzLr+0vtPC/N1z0x6V9I1H5EPneb5SeZ/f2jP5114bEKhe8b3OPE4Z1+W0rWFhMht4/O4k2jfj1xzT6KK5DsDr1oAA6DFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWH4o8aeHfBlolz4n1aDT45M+WHyzyY67UUFmxkdBW5Xnni2w1DRPiLaeN7fQn1+zi037DNb2wDXVpiRn86JDw+Q20gENwMZ5FAG/oXxC8LeJtDvdX0PV47uysFL3TrG6tEApYkoyhugPbnHFaPh7xHpPivRIdX8P3i3ljOWCSqrLkqSCCrAEHI7iucu9f0DxP8N/EmqeHZIpS2n3EVyRF5cyOsTfJIpAYMMng+vHWuS+ETHwpdaLozkjT/E+hWup2Wei3SW8a3CD/AHl2Sf8AfVAHpuveKdG8M/YRrl6tqdQuVtLUFGcyyt0UBQcfU8Duafr/AIk0fwtppv8AxDqMGn2udokmbG4+gHUn2HNeN+P528T+Km1vh9N0PWrDR7HPIac3Mb3Mg9wVSPP+y1dfZ6ZaeKvjhr1zrcK3SeGbe1t9Pt5lzHG8yGR5Qp4LdFB/2fYUAbPhr4reCfF9+LHw/wCILe5u2zsgdHhd8DJ2iRVLcemao6t8bvh7oer3Wl6p4g8i8tJTFPF9iuG2ODgjKxkH8DU3xZ0Cx1P4c6vfyQol/pNnJfWN4igS28kKmRSrdRyuPxrI+I1/JqnwY0nUJwBLd3GlzuB0DNNEx/nQBuaT8WfBOuaNquq6XrXn2Wjokl9L9kmXyVbdtOGQFs7W+6D0qtovxq8AeIdattJ0fX/tF7dPshi+x3Cbm64yyADp3NTfGL/kjnib/rxb+Yqr4V+Jf9s3un6V/wAIV4wsPNQJ9svtK8q3TCZyz7jgHGAcdxQB39FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVx/ijWvHOkauP+Eb8JWniHTmiHA1JbWZJMnOd4KkYxjHvXYUUAeY6T4W8T3tr411zXrG1sNT8Q2AtbfS7ScSBNkTqpkk4VnJfGegA/J+reDNbn+Dfhu20uNbfxV4dtLOa0UupC3EUSq8RbO0hhuU84PHNel0UAea6r4AvbL4ZeHfD2jQ/bLmx1Oyurt/MVTIVmEk0pLEZySx9frVzxV4c8Sab4u/4TDwEtrc3k9utrqWl3chjS8RCSjq/wDDIuSMnjB/PvqKAPK9RtfiJ8RrQ6JrmgWvg/RJ2Av3/tFbu4uIsgmOPyxtXd0JPbP0PR/Efw9e6z4Hi0vQbQSyRXtm6wh1QLHHMjNyxA4VTx7V2NFAHL/ErR77xB8Ndd0rSIPtF7d2rRwxb1Xe2RxliAPxNYei+KviK13p9lqHww+x2heOGe7/AOEgt5PJTIDSbAMtgZO0cnGK9EooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigD/9k=)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_EPJmdxHA8zF"
      },
      "source": [
        "#PCA analysis\n",
        "def PCA_anal(x, n_comp):\n",
        "  pca = PCA(n_components= n_comp)\n",
        "  printcipalComponents = pca.fit_transform(x)\n",
        "  \n",
        "  return x"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FZsUl0IVBif7"
      },
      "source": [
        "**PCA**          \n",
        "PCA is a method of finding the principal components of data. However, PCA does not analyze the components of each data, but rather analyzes the principal components of this distribution when several data are collected to form a single distribution. In other words, it is a technique that converts samples from high-dimensional space into low-dimensional space by finding axes that are orthogonal to each other while preserving variance as much as possible. The principal component refers to the direction vector in which the variance of data is greatest in that direction.\n",
        "\n",
        "'Explained Variance Ratio' refers to the variance ratio result of projection on the each principal component vector. When looking at the output of the variance ratio, 0.54819966, 0.0447017, 0.03638499 are obtained. This means that 55% of the variance in the original dataset lies on the first principal component axis, and about 4.5% lies on the second principal component axis. As a result of reducing the dimension of the data set by using 20 principal components, only about 72% of data is used, and 28% is lost."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V8ay34pv0m6i"
      },
      "source": [
        "# 1-1. MLP Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D00VFC4ctnba"
      },
      "source": [
        "#Dataset\n",
        "class getDataset(Dataset):\n",
        "\n",
        "  def __init__(self, x, y):\n",
        "    #Normalization\n",
        "    mu = np.mean(x, axis = 0)\n",
        "    std = np.std(x, axis = 0)\n",
        "    x = (x - mu) / std\n",
        "    \n",
        "    self.x_train = torch.tensor(x, dtype= torch.float32)\n",
        "    self.y_train = torch.tensor(y)\n",
        "\n",
        "  def __len__(self):\n",
        "    return len(self.y_train)\n",
        "\n",
        "  def __getitem__(self, idx):\n",
        "    return self.x_train[idx], self.y_train[idx]\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iA4dD9e3gfAU"
      },
      "source": [
        "#model\n",
        "class Net(nn.Module):\n",
        "\n",
        "  def __init__(self):\n",
        "    super(Net, self).__init__()\n",
        "    fc1 = nn.Linear(140, 140)\n",
        "    fc2 = nn.Linear(23, 23)\n",
        "    fc3 = nn.Linear(140, 3)\n",
        "    fc4 = nn.Linear(140,23)\n",
        "    fc5 = nn.Linear(23, 3)\n",
        "    dropout = torch.nn.Dropout(p=0.2)\n",
        "    self.KNlayer = nn.Sequential(fc1,\n",
        "                                nn.ReLU(),\n",
        "                                fc3,)\n",
        "    \n",
        "    self.KNlayer2 = nn.Sequential(fc2,\n",
        "                                nn.ReLU(),\n",
        "                                fc4 )\n",
        "  def forward(self, x):\n",
        "     output = self.KNlayer(x)\n",
        "     #output = self.KNlayer2(x)\n",
        "  \n",
        "     return F.log_softmax(output)\n",
        "  \n",
        "def weight_init(m):\n",
        "    if type(m) == nn.Linear:\n",
        "      nn.init.kaiming_uniform_(m.weight)\n",
        "      nn.init.constant_(m.bias, 0.01)\n",
        "\n",
        "\n",
        "model = Net()\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1DQdf7AEpamN"
      },
      "source": [
        "#Loss function and optimizer\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.Adam(model.parameters(), lr = 0.0001)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RWan_7Gi2DEC"
      },
      "source": [
        "#train \n",
        "num_epochs = 500\n",
        "avg_val_acc = 0\n",
        "avg_val_loss = 0\n",
        "avg_val_list = []\n",
        "train_batch = 128\n",
        "valid_batch = 16\n",
        "patience = 100  # early stopping patience\n",
        "\n",
        "for i in range(10):\n",
        "  print('//////class {0}/////////'.format(i))\n",
        "  \n",
        "  train_dataset = getDataset(x_train_list[i], y_train_list[i])\n",
        "  validation_dataset = getDataset(x_valid_list[i], y_valid_list[i])\n",
        "\n",
        "  train_loader = torch.utils.data.DataLoader(train_dataset, batch_size= train_batch, shuffle = True)\n",
        "  val_loader = torch.utils.data.DataLoader(validation_dataset, batch_size= valid_batch, shuffle = False)\n",
        "  \n",
        "  num_train =len(train_dataset)\n",
        "  num_valid = len(validation_dataset)\n",
        "  \n",
        "  early_stopping = EarlyStopping(patience=patience, verbose=True)\n",
        "  for epoch in range(num_epochs):\n",
        "\n",
        "    model.train()\n",
        "    total_loss = 0\n",
        "    running_correct = 0\n",
        "    val_running_loss = 0\n",
        "    val_running_correct = 0\n",
        "\n",
        "    for train_x, train_y in train_loader: \n",
        "      output = model(train_x)\n",
        "      loss = criterion(output, train_y)\n",
        "      \n",
        "      _, predict = torch.max(output, 1)\n",
        "\n",
        "      optimizer.zero_grad()\n",
        "      loss.backward()\n",
        "\n",
        "      optimizer.step()\n",
        "\n",
        "      total_loss += loss.item()\n",
        "      running_correct += torch.sum(predict == train_y.data) \n",
        "    epoch_loss = total_loss / (num_train / train_batch)\n",
        "    epoch_acc = running_correct.double() / num_train * 100\n",
        "    \n",
        "    #validation\n",
        "    model.eval()\n",
        "    with torch.no_grad():\n",
        "      for val_x, val_y in val_loader:\n",
        "        val_output = model(val_x)\n",
        "        val_loss = criterion(val_output, val_y)\n",
        "        _, val_predict = torch.max(val_output, 1)\n",
        "\n",
        "        val_running_loss += val_loss.item()\n",
        "        val_running_correct += torch.sum(val_predict == val_y.data) \n",
        "\n",
        "      val_epoch_loss = val_running_loss / (num_valid / valid_batch)\n",
        "      val_epoch_acc = val_running_correct.double() / num_valid * 100\n",
        "\n",
        "    #early stopping\n",
        "    '''\n",
        "      if epoch >= 40:\n",
        "        early_stopping(val_loss, model)\n",
        "\n",
        "        if early_stopping.early_stop:\n",
        "            print(\"Early stopping\")\n",
        "            avg_val_list.append(val_epoch_acc.float().tolist())\n",
        "            avg_val_loss += val_epoch_loss\n",
        "            avg_val_acc += val_epoch_acc\n",
        "            break\n",
        "    '''\n",
        "    if ((epoch + 1) % 100 == 0):\n",
        "\n",
        "      print('Epoch {}/{}'.format(epoch+1, num_epochs))\n",
        "      print(\"training loss: {:.5f}, acc: {:5f}\".format(epoch_loss, epoch_acc))\n",
        "      print(\"validation loss: {:.5f}, acc: {:5f}\".format(val_epoch_loss, val_epoch_acc))\n",
        "      print(\"===================================================\")\n",
        "\n",
        "    if (epoch == (num_epochs - 1)):\n",
        "      avg_val_list.append(val_epoch_acc.float().tolist())\n",
        "      avg_val_loss += val_epoch_loss\n",
        "      avg_val_acc += val_epoch_acc\n",
        "  model.apply(weight_init)\n",
        "avg_val_loss /= 10\n",
        "avg_val_acc /= 10\n",
        "\n",
        "print(avg_val_list)   \n",
        "print(\"avg validation loss: {:.5f}, acc: {:5f}\".format(avg_val_loss, avg_val_acc))\n",
        "        "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GCYiIn5Cu8Zj"
      },
      "source": [
        "**Dataset setting & Estabilish model architecture**           \n",
        "We made a simple classifier that predicts labels by learning based on multi-layer perceptron and softmax. The code was mainly implemented through the pytorch library. The dataset was converted to a normal distribution to make it the same as the 'Standardscaler()' function used in other models. Designing the model's architecture was very tricky process. Due to the  data is very small, if the layers are deeply stacked, it is difficult to learn and it requires a lot of computational cost. So, the model was composed of two layers, and nonlinearity was given by adding an activation function, ReLU function, between them.\n",
        "\n",
        "**Experiments Result**       \n",
        "When using various preprocessed datasets and performing 10-fold cross validation, the average accuracy was 51%. It might be thought that we've got the answer half right, but in fact, the MCI group accounts for about 50% of the entire dataset, so it's hardly learned. Even if all are predicted by the MCI group, there will be a probability of 50% or more. I tried training several times while adjusting various hyperparameters (batch size, learning rate, etc.), but did not get any meaningful results."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S3PSf1AQlOY-"
      },
      "source": [
        "# 1-2. LIGHTGBM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DfMT-9gJlk6E"
      },
      "source": [
        "params = {\n",
        "    \"n_estimators\" : 600,\n",
        "    \"num_leaves\" : 64,\n",
        "    \"subsample\" : 0.8,\n",
        "    \"colsample_bytree\" : 0.7,\n",
        "    \"objective\" : 'multiclass'\n",
        "}\n",
        "params2 = {\n",
        "    \"n_estimators\" : 500,\n",
        "    \"learning_rate\" : 0.3,\n",
        "    \"max_depth\" : 10,\n",
        "    \"alpha\" : 0\n",
        "}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VNRCph-RkzbD"
      },
      "source": [
        "models = []\n",
        "best_iter = 0\n",
        "_y_predict = np.zeros(x_train.shape[0])\n",
        "pred_fold = 0\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x_BbZP8kOn2R",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d70f9d1e-67b1-4ee7-c20e-cb74854eac16"
      },
      "source": [
        "skf = StratifiedKFold(10,shuffle=True, random_state =41)\n",
        "acc_avg = 0\n",
        "for train_index, valid_index in skf.split(x, y):\n",
        "  x_train, x_valid = x.iloc[train_index, :], x.iloc[valid_index, :]\n",
        "  y_train, y_valid= y.iloc[train_index], y.iloc[valid_index]\n",
        "  \n",
        "  \n",
        "  #x_train, y_train = Resampling(x_train, y_train, 'SMOTE')\n",
        "  x_train = PCA_anal(x_train, 20)\n",
        "\n",
        "  scaler = StandardScaler()\n",
        "  scaler.fit(x_train)\n",
        "  x_train = scaler.transform(x_train)\n",
        "  x_valid = scaler.transform(x_valid)\n",
        "\n",
        "  \n",
        "  #model = XGBClassifier(**params2)\n",
        "  model = LGBMClassifier(**params)\n",
        "  model.fit(x_train, y_train, verbose=False)\n",
        "  \n",
        "  pred_fold += model.predict_proba(x_valid[:159])\n",
        "  pred = np.argmax(pred_fold,axis=1)\n",
        "  print('Prediction :', Counter(pred))\n",
        "  print('True Value : ', Counter(y_valid))\n",
        "  y_valid = y_valid[:159]\n",
        "  \n",
        "  \n",
        "  accuracy = accuracy_score(y_valid,pred)\n",
        "  acc_avg += accuracy\n",
        "  precision = precision_score(y_valid,pred,average='micro')\n",
        "  recall = recall_score(y_valid,pred,average='micro')\n",
        "  mse = mean_squared_error(y_valid, pred)\n",
        "\n",
        "\n",
        "  print('accuracy : %f' %(accuracy * 100))\n",
        "  print('Mean squared error: ', mse)\n",
        "  print('Precision: ', precision)\n",
        "  print('Recall: ', recall)\n",
        "  print(\"================================\")\n",
        "\n",
        "print('Average Accuracy :', (acc_avg / 10) * 100 )\n",
        "print(\"The End\") "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Prediction : Counter({1: 110, 0: 32, 2: 17})\n",
            "True Value :  Counter({1: 81, 0: 49, 2: 31})\n",
            "accuracy : 55.974843\n",
            "Mean squared error:  0.4591194968553459\n",
            "Precision:  0.559748427672956\n",
            "Recall:  0.559748427672956\n",
            "================================\n",
            "Prediction : Counter({1: 120, 2: 20, 0: 19})\n",
            "True Value :  Counter({1: 81, 0: 49, 2: 31})\n",
            "accuracy : 50.314465\n",
            "Mean squared error:  0.5345911949685535\n",
            "Precision:  0.5031446540880503\n",
            "Recall:  0.5031446540880503\n",
            "================================\n",
            "Prediction : Counter({1: 129, 0: 23, 2: 7})\n",
            "True Value :  Counter({1: 81, 0: 49, 2: 31})\n",
            "accuracy : 46.540881\n",
            "Mean squared error:  0.6289308176100629\n",
            "Precision:  0.46540880503144655\n",
            "Recall:  0.46540880503144655\n",
            "================================\n",
            "Prediction : Counter({1: 128, 0: 24, 2: 7})\n",
            "True Value :  Counter({1: 82, 0: 49, 2: 30})\n",
            "accuracy : 49.685535\n",
            "Mean squared error:  0.6163522012578616\n",
            "Precision:  0.4968553459119497\n",
            "Recall:  0.4968553459119497\n",
            "================================\n",
            "Prediction : Counter({1: 132, 0: 23, 2: 4})\n",
            "True Value :  Counter({1: 82, 0: 49, 2: 30})\n",
            "accuracy : 53.459119\n",
            "Mean squared error:  0.559748427672956\n",
            "Precision:  0.5345911949685535\n",
            "Recall:  0.5345911949685535\n",
            "================================\n",
            "Prediction : Counter({1: 136, 0: 21, 2: 2})\n",
            "True Value :  Counter({1: 82, 0: 48, 2: 31})\n",
            "accuracy : 44.654088\n",
            "Mean squared error:  0.6666666666666666\n",
            "Precision:  0.44654088050314467\n",
            "Recall:  0.44654088050314467\n",
            "================================\n",
            "Prediction : Counter({1: 147, 0: 12})\n",
            "True Value :  Counter({1: 82, 0: 48, 2: 31})\n",
            "accuracy : 50.314465\n",
            "Mean squared error:  0.5723270440251572\n",
            "Precision:  0.5031446540880503\n",
            "Recall:  0.5031446540880503\n",
            "================================\n",
            "Prediction : Counter({1: 152, 0: 6, 2: 1})\n",
            "True Value :  Counter({1: 81, 0: 48, 2: 31})\n",
            "accuracy : 52.201258\n",
            "Mean squared error:  0.4779874213836478\n",
            "Precision:  0.5220125786163522\n",
            "Recall:  0.5220125786163522\n",
            "================================\n",
            "Prediction : Counter({1: 147, 0: 9, 2: 3})\n",
            "True Value :  Counter({1: 81, 0: 48, 2: 31})\n",
            "accuracy : 52.830189\n",
            "Mean squared error:  0.5094339622641509\n",
            "Precision:  0.5283018867924528\n",
            "Recall:  0.5283018867924528\n",
            "================================\n",
            "Prediction : Counter({1: 149, 0: 8, 2: 2})\n",
            "True Value :  Counter({1: 81, 0: 48, 2: 31})\n",
            "accuracy : 55.345912\n",
            "Mean squared error:  0.46540880503144655\n",
            "Precision:  0.5534591194968553\n",
            "Recall:  0.5534591194968553\n",
            "================================\n",
            "Average Accuracy : 51.132075471698116\n",
            "The End\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FvuTKNRJDdsS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c6dcb70-efbd-4e2a-acf9-1deaf02c6dd4"
      },
      "source": [
        "save = pickle.dumps(model)\n",
        "model2 = pickle.loads(save)\n",
        "model2.predict(x_train[0:1])\n",
        "\n",
        "joblib.dump(model2, 'cls_model.pkl') "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['cls_model.pkl']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 116
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "inNHmNVPYYvu"
      },
      "source": [
        "![ff.JPG](data:image/jpeg;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "axRr1FW5YBND"
      },
      "source": [
        "**Experiments Result**          \n",
        "As a result of analyzing the data, the above table is a drop of data and features that are considered to be outlier or have a small correlation. It was predicted that removing outliers or features with little correlation would have more effect on classification task, but according to the experimental results, the baseline model performed best and the preprocessed dataset did not perform better than that.         \n",
        "\n",
        "Also, various resampling strategies were applied. The undersampling and oversampling techniques were applied to the preprocessed data sets and compared the results. It is a value obtained by calculating the average of validation accuracies from 10-fold cross validation, and the accuracy was repeatedly measured several times. Before the experiment, since there was a lot of overlapping data, it was expected that TomekLinks resampling would show good performance, because it distinguishes the boundary, but overall most of the strategies did not show a big improvement.\n",
        "\n",
        "It can be said that the above result was also practically not learned. The best average accuracy is 51%. When PCA analysis is used, it tends to improve by 1%, but there is no significant effect. We compared the number of predicted labels and actual labels by class. The actual label is evaluated according to the ratio using the StratifiedKFold() function, while the predicted label only shows the MCI, which is the majority group, as a predicted value.\n",
        "\n",
        "**LightGBM**      \n",
        "LightGBM is the Gradient Boosting framework and learning algorithm based on tree. LightGBM expands the tree vertically, while other algorithms expand horizontally. i.e. LightGBM is leaf-wise while other algorithms are level-wise. In this decision tree model, algorithms such as Xgboost and Randomforest also show good performance, but as a result of experiments, LightGBM has the fastest learning speed and good performance.\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JNrdLY6CNtuZ"
      },
      "source": [
        "# 1-3. SVM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mvAQXPFBCHYa"
      },
      "source": [
        "skf = StratifiedKFold(10,shuffle=False, random_state =41)\n",
        "pred_fold = 0\n",
        "acc_avg = 0\n",
        "for train_index, valid_index in skf.split(x, y):\n",
        "  x_train, x_valid = x.iloc[train_index, :].values, x.iloc[valid_index, :].values\n",
        "  y_train, y_valid= y.iloc[train_index].values, y.iloc[valid_index].values\n",
        "  \n",
        "  #x_train, y_train = Resampling(x_train, y_train, 'SMOTE')\n",
        "  #x_train = PCA_anal(x_train, 20)\n",
        "\n",
        "  scaler = StandardScaler()\n",
        "  scaler.fit(x_train)\n",
        "  x_train = scaler.transform(x_train)\n",
        "  x_valid = scaler.transform(x_valid)\n",
        "  \n",
        "  #PCA analysis\n",
        "  pca = PCA(n_components=25)\n",
        "  printcipalComponents = pca.fit_transform(x_train)\n",
        "  \n",
        "\n",
        "  estimator = LinearSVC()\n",
        "  n_estimators = 100\n",
        "  n_jobs = 2\n",
        "  model = BaggingClassifier(base_estimator=estimator,\n",
        "                          n_estimators=n_estimators,\n",
        "                          max_samples=1./n_estimators,\n",
        "                          n_jobs=n_jobs)\n",
        "  model.fit(x_train,y_train)\n",
        "  pred_fold += model.predict_proba(x_valid[:159])\n",
        "  pred = np.argmax(pred_fold,axis=1)\n",
        "  y_valid = y_valid[:159]\n",
        "  accuracy = accuracy_score(y_valid,pred)\n",
        "  acc_avg += accuracy\n",
        "  print('Accuracy :',accuracy)\n",
        "print('Average Accuracy :', ( acc_avg / 10) * 100)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5GCjFDGcR5Q6"
      },
      "source": [
        "![aaaaaa.JPG](data:image/jpeg;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "31H-jwPSNMlH"
      },
      "source": [
        "**SVM classification**       \n",
        "Support Vector Machine (SVM) is an algorithm that is evaluated to perform much better than other machine learning algorithms. However, SVM has very high time/space complexity. To overcome this, a method of maintaining the SVM perfoemance and lowering the cost by using an ensemble technique has been proposed. \n",
        "Bagging is one of the ensemble techniques, which creates several small data sets by reconstruction extraction from training data, and creates a classifier with each data set. Then when we predict test label, we  uses a voting to decide class and label.\n",
        "\n",
        "I tried using the support vector machine and the ensemble method, which are known to have good performance, but the results were not different from other models."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8n4RZJesxoor"
      },
      "source": [
        "# 2-1. Linear Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6M3xOz7Q9_Q4"
      },
      "source": [
        "data_2 = data.iloc[:,:4]\n",
        "\n",
        "x_reg = data_1.iloc[:,1:]\n",
        "y_reg = data_2.iloc[:,1:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CnlQSarnxX-6",
        "scrolled": false,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 476
        },
        "outputId": "fcfbcca5-4ba1-4d41-b9be-9e0bc7fa8274"
      },
      "source": [
        "kf = KFold(10,shuffle=True, random_state =1)\n",
        "mse_avg = 0\n",
        "for train_index, valid_index in kf.split(x_reg):\n",
        "    x_train, x_valid = x_reg.iloc[train_index, :].values, x_reg.iloc[valid_index, :].values\n",
        "    y_train, y_valid= y_reg.iloc[train_index].values, y_reg.iloc[valid_index].values\n",
        "\n",
        "    scaler = StandardScaler()\n",
        "    poly_Reg = PolynomialFeatures(degree=2, include_bias=True)\n",
        "    # Fit on training set only.\n",
        "    \n",
        "    # Apply transform to both the training set and the test set.\n",
        "    #x_train = poly_Reg.fit_transform(x_train)\n",
        "    #x_valid = poly_Reg.fit_transform(x_valid)\n",
        "    scaler.fit(x_train)\n",
        "    x_train = scaler.transform(x_train)\n",
        "    x_valid = scaler.transform(x_valid) \n",
        "    \n",
        "\n",
        "    model = LinearRegression()\n",
        "    #model = LogisticRegression()\n",
        "\n",
        "    model.fit(x_train,y_train)\n",
        "    predictions = model.predict(x_valid)\n",
        "    \n",
        "    \n",
        "    MSE = mean_squared_error(y_valid,predictions)\n",
        "    print('MSE :', MSE)\n",
        "    mse_avg += MSE\n",
        "    plt.scatter(y_valid, predictions)\n",
        "    plt.xlabel('Y Test')\n",
        "    plt.ylabel('Predicted Y')\n",
        "    \n",
        "print('Average MSE : ', (mse_avg / 10) )  "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MSE : 32.6418137687326\n",
            "MSE : 27.00452968270636\n",
            "MSE : 29.712508268292336\n",
            "MSE : 32.879317150191426\n",
            "MSE : 25.049620274896807\n",
            "MSE : 24.8083195667836\n",
            "MSE : 23.517832369142287\n",
            "MSE : 28.523998886590245\n",
            "MSE : 34.84505960991262\n",
            "MSE : 25.123767456121744\n",
            "Average MSE :  28.410676703337003\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6h_b_ZyywA37",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b7202b93-2e79-4e35-cec6-6a2d4583fe41"
      },
      "source": [
        "save = pickle.dumps(model)\n",
        "model2 = pickle.loads(save)\n",
        "model2.predict(x_train[0:1])\n",
        "model2\n",
        "joblib.dump(model2, 'reg_model.pkl') "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['reg_model.pkl']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 118
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E-l09rcN0wEt"
      },
      "source": [
        "# 2-2. Neural Network Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EZ4gSc9_9_Q9"
      },
      "source": [
        "epoch = 500\n",
        "learning_rate = 0.00003\n",
        "batch_size = 64"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yTfW70o0vk-b"
      },
      "source": [
        "class AD_Prediction(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(AD_Prediction, self).__init__()\n",
        "        \n",
        "        self.layer1 = nn.Sequential(nn.Linear(140, 128),\n",
        "                                    nn.BatchNorm1d(128),\n",
        "                                    nn.ReLU(),\n",
        "                                    nn.Linear(128, 64),\n",
        "                                    nn.BatchNorm1d(64),\n",
        "                                    nn.ReLU())\n",
        "        self.layer2 = nn.Sequential(nn.Linear(64, 32),\n",
        "                                    nn.ReLU(),\n",
        "                                    nn.Linear(32, 3) )                                  \n",
        "                                    \n",
        "    \n",
        "    def forward(self, x):\n",
        "        out = self.layer1(x)\n",
        "        out = self.layer2(out)\n",
        "    \n",
        "        return out\n",
        "\n",
        "    \n",
        "def weight_init(m):\n",
        "    if type(m) == nn.Linear:\n",
        "        nn.init.kaiming_uniform_(m.weight)\n",
        "        nn.init.constant_(m.bias, 0.01)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FP9EVLrgvssw"
      },
      "source": [
        "model = AD_Prediction()\n",
        "criterion = nn.MSELoss()\n",
        "optimizer = optim.SGD(model.parameters(), lr=learning_rate)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xH3Yw3qJv5OV"
      },
      "source": [
        "kf = KFold(10,shuffle=True, random_state =41)\n",
        "\n",
        "for train_index, valid_index in kf.split(x_reg):\n",
        "  x_train, x_valid = x_reg.iloc[train_index, :].values, x_reg.iloc[valid_index, :].values\n",
        "  y_train, y_valid= y_reg.iloc[train_index].values, y_reg.iloc[valid_index].values\n",
        "\n",
        "  scaler = StandardScaler()\n",
        "  # Fit on training set only.\n",
        "\n",
        "  # Apply transform to both the training set and the test set.\n",
        "  scaler.fit(x_train)\n",
        "  x_train = scaler.transform(x_train)\n",
        "  x_valid = scaler.transform(x_valid) \n",
        "\n",
        "  train_loader = torch.utils.data.DataLoader(x_train, batch_size=batch_size, shuffle = True)\n",
        "  train_loader_y = torch.utils.data.DataLoader(y_train, batch_size=batch_size, shuffle = True)\n",
        "\n",
        "  val_loader = torch.utils.data.DataLoader(x_valid, batch_size=1, shuffle = False)\n",
        "  val_loader_y = torch.utils.data.DataLoader(y_valid, batch_size=1, shuffle = False)\n",
        "      \n",
        "\n",
        "  list_train_loss = []\n",
        "  list_valid_loss = []\n",
        "\n",
        "  model.apply(weight_init)\n",
        "  \n",
        "  for i in range(epoch):            \n",
        "      total_train_loss = 0\n",
        "      total_valid_loss = 0\n",
        "      \n",
        "      #train\n",
        "      for x, y in zip(train_loader, train_loader_y):\n",
        "          x = x.float()\n",
        "          y = y.float()\n",
        "          output = model(x)\n",
        "          loss = criterion(output, y)\n",
        "\n",
        "          optimizer.zero_grad()\n",
        "          loss.backward()\n",
        "          optimizer.step()\n",
        "          \n",
        "          total_train_loss += loss.item()\n",
        "          \n",
        "      \n",
        "      #validation\n",
        "      model.eval()\n",
        "      with torch.no_grad():\n",
        "          for x, y in zip(val_loader, val_loader_y):\n",
        "              x = x.float()\n",
        "              y = y.float()\n",
        "              val_output = model(x)\n",
        "              val_loss = criterion(val_output, y)\n",
        "  \n",
        "              total_valid_loss += val_loss.item()\n",
        "\n",
        "      if ((i + 1) % 10 == 0):\n",
        "        print('Epoch',(i+1))\n",
        "        print(\"training MSE loss: {:.5f}\".format((total_train_loss / (len(x_train)/batch_size + 1))))\n",
        "        print(\"validation MSE loss: {:.5f}\".format((total_valid_loss / len(x_valid))))\n",
        "        print(\"===================================================\")\n",
        "      \n",
        "      list_train_loss.append(total_train_loss / (len(x_train)/batch_size + 1))\n",
        "      list_valid_loss.append(total_valid_loss / len(x_valid))\n",
        "  \n",
        "  # plot epoch-loss function\n",
        "  print(\"\\n\\n\\n===================================================\")\n",
        "  print(\"===================================================\")\n",
        "  print(\"K-FOLD CHANGE\")\n",
        "  \n",
        "  plt.plot(np.arange(0,epoch,1), np.array(list_train_loss),color='red')\n",
        "  plt.plot(np.arange(0,epoch,1), np.array(list_valid_loss),color='blue')\n",
        "  plt.xlabel('epoch')\n",
        "  plt.ylabel('loss')\n",
        "  plt.show()\n",
        "  \n",
        "  print(\"===================================================\")\n",
        "  print(\"===================================================\\n\\n\\n\")\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GF7KyYmt1kEb"
      },
      "source": [
        "# 2-3. K-NN Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3vYh4Equv8sq"
      },
      "source": [
        "kf = KFold(10,shuffle=True, random_state =1)\n",
        "mse_avg = 0\n",
        "for train_index, valid_index in kf.split(x_reg):\n",
        "  x_train, x_valid = x_reg.iloc[train_index, :].values, x_reg.iloc[valid_index, :].values\n",
        "  y_train, y_valid= y_reg.iloc[train_index].values, y_reg.iloc[valid_index].values\n",
        "\n",
        "  scaler = StandardScaler()\n",
        "  scaler.fit(x_train)\n",
        "  x_train = scaler.transform(x_train)\n",
        "  x_valid = scaler.transform(x_valid)\n",
        "\n",
        "  regressor = KNeighborsRegressor(n_neighbors = 5, weights = \"distance\")\n",
        "  regressor.fit(x_train, y_train)\n",
        "\n",
        "  predictions = regressor.predict(x_valid)\n",
        "    \n",
        "  MSE = mean_squared_error(y_valid,predictions)\n",
        "  print('MSE :', MSE)\n",
        "  mse_avg += MSE\n",
        "print('Average MSE : ', (mse_avg / 10) ) "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vqwLmSCQdAo9"
      },
      "source": [
        "**Regression Results**         \n",
        "\n",
        "![sdsdsdsddddd.JPG](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAeAB4AAD/4RD6RXhpZgAATU0AKgAAAAgABAE7AAIAAAAQAAAISodpAAQAAAABAAAIWpydAAEAAAAgAAAQ0uocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEd5ZW9uZ055ZW9uIEtpbQAABZADAAIAAAAUAAAQqJAEAAIAAAAUAAAQvJKRAAIAAAADNTkAAJKSAAIAAAADNTkAAOocAAcAAAgMAAAInAAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADIwMjA6MTE6MDEgMTg6MDk6MDUAMjAyMDoxMTowMSAxODowOTowNQAAAEcAeQBlAG8AbgBnAE4AeQBlAG8AbgAgAEsAaQBtAAAA/+ELImh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMjAtMTEtMDFUMTg6MDk6MDUuNTg4PC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPkd5ZW9uZ055ZW9uIEtpbTwvcmRmOmxpPjwvcmRmOlNlcT4NCgkJCTwvZGM6Y3JlYXRvcj48L3JkZjpEZXNjcmlwdGlvbj48L3JkZjpSREY+PC94OnhtcG1ldGE+DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgIDw/eHBhY2tldCBlbmQ9J3cnPz7/2wBDAAcFBQYFBAcGBQYIBwcIChELCgkJChUPEAwRGBUaGRgVGBcbHichGx0lHRcYIi4iJSgpKywrGiAvMy8qMicqKyr/2wBDAQcICAoJChQLCxQqHBgcKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKir/wAARCAHJA2ADASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwD6RoorK8Ra3/YGlfbfs/2j94E2b9vXvnBqZSUIuUthNpK7NWivP/8AhaP/AFCP/Jn/AOwo/wCFo/8AUI/8mf8A7CuX67h/5vwZl7en3PQKK8//AOFo/wDUI/8AJn/7Cj/haP8A1CP/ACZ/+wo+u4f+b8GHt6fc9Aorz/8A4Wj/ANQj/wAmf/sKP+Fo/wDUI/8AJn/7Cj67h/5vwYe3p9z0CivP/wDhaP8A1CP/ACZ/+wo/4Wj/ANQj/wAmf/sKPruH/m/Bh7en3PQKK8//AOFo/wDUI/8AJn/7Cj/haP8A1CP/ACZ/+wo+u4f+b8GHt6fc9Aorz/8A4Wj/ANQj/wAmf/sKP+Fo/wDUI/8AJn/7Cj67h/5vwYe3p9z0CivP/wDhaP8A1CP/ACZ/+wo/4Wj/ANQj/wAmf/sKPruH/m/Bh7en3PQKK8//AOFo/wDUI/8AJn/7Cj/haP8A1CP/ACZ/+wo+u4f+b8GHt6fc9Aorz/8A4Wj/ANQj/wAmf/sKP+Fo/wDUI/8AJn/7Cj67h/5vwYe3p9z0CivP/wDhaP8A1CP/ACZ/+wo/4Wj/ANQj/wAmf/sKPruH/m/Bh7en3PQKK8//AOFo/wDUI/8AJn/7Cj/haP8A1CP/ACZ/+wo+u4f+b8GHt6fc9Aorz/8A4Wj/ANQj/wAmf/sKP+Fo/wDUI/8AJn/7Cj67h/5vwYe3p9z0CivP/wDhaP8A1CP/ACZ/+wq3dfEP7Np9jc/2Zu+1o7bftGNm1yvXbz0zTWMoP7X4MPbU+52tFef/APC0f+oR/wCTP/2FH/C0f+oR/wCTP/2FL67h/wCb8GHt6fc9Aorz/wD4Wj/1CP8AyZ/+wo/4Wj/1CP8AyZ/+wo+u4f8Am/Bh7en3PQKK8/8A+Fo/9Qj/AMmf/sKP+Fo/9Qj/AMmf/sKPruH/AJvwYe3p9z0CivP/APhaP/UI/wDJn/7Cj/haP/UI/wDJn/7Cj67h/wCb8GHt6fc9Aorz/wD4Wj/1CP8AyZ/+wo/4Wj/1CP8AyZ/+wo+u4f8Am/Bh7en3PQKK8/8A+Fo/9Qj/AMmf/sKP+Fo/9Qj/AMmf/sKPruH/AJvwYe3p9z0CivP/APhaP/UI/wDJn/7Cj/haP/UI/wDJn/7Cj67h/wCb8GHt6fc9Aorz/wD4Wj/1CP8AyZ/+wo/4Wj/1CP8AyZ/+wo+u4f8Am/Bh7en3PQKK8/8A+Fo/9Qj/AMmf/sKP+Fo/9Qj/AMmf/sKPruH/AJvwYe3p9z0CivP/APhaP/UI/wDJn/7Cj/haP/UI/wDJn/7Cj67h/wCb8GHt6fc9Aorz/wD4Wj/1CP8AyZ/+wo/4Wj/1CP8AyZ/+wo+u4f8Am/Bh7en3PQKK8/8A+Fo/9Qj/AMmf/sKP+Fo/9Qj/AMmf/sKPruH/AJvwYe3p9z0CivP/APhaP/UI/wDJn/7Creq/EP8AszVJ7P8AszzfJbbv+0Yzx6baf1yha/N+DD21O17na0V5/wD8LR/6hH/kz/8AYUf8LR/6hH/kz/8AYUvruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKP+Fo/9Qj/yZ/8AsKPruH/m/Bh7en3PQKK8/wD+Fo/9Qj/yZ/8AsKt3fxD+y2Vlcf2Zu+1xs+37RjbhyuPu89M01jKD15vwYe2p9ztaK8//AOFo/wDUI/8AJn/7Cj/haP8A1CP/ACZ/+wpfXcP/ADfgw9vT7noFFef/APC0f+oR/wCTP/2FH/C0f+oR/wCTP/2FH13D/wA34MPb0+56BRXn/wDwtH/qEf8Akz/9hR/wtH/qEf8Akz/9hR9dw/8AN+DD29PuegUV5/8A8LR/6hH/AJM//YUf8LR/6hH/AJM//YUfXcP/ADfgw9vT7noFFef/APC0f+oR/wCTP/2FH/C0f+oR/wCTP/2FH13D/wA34MPb0+56BRXn/wDwtH/qEf8Akz/9hR/wtH/qEf8Akz/9hR9dw/8AN+DD29PuegUV5/8A8LR/6hH/AJM//YUf8LR/6hH/AJM//YUfXcP/ADfgw9vT7noFFef/APC0f+oR/wCTP/2FH/C0f+oR/wCTP/2FH13D/wA34MPb0+56BRXn/wDwtH/qEf8Akz/9hR/wtH/qEf8Akz/9hR9dw/8AN+DD29PuegUV5/8A8LR/6hH/AJM//YUf8LR/6hH/AJM//YUfXcP/ADfgw9vT7noFFef/APC0f+oR/wCTP/2FH/C0f+oR/wCTP/2FH13D/wA34MPb0+56BRXn/wDwtH/qEf8Akz/9hR/wtH/qEf8Akz/9hR9dw/8AN+DD29PuegUV5/8A8LR/6hH/AJM//YUf8LR/6hH/AJM//YUfXcP/ADfgw9vT7noFFc74X8Wf8JJNcJ9i+zeQqnPm792c+w9K6KumnUjUjzR2NIyUldBRRRVlBXK/ET/kVT/13T+tdVXK/ET/AJFU/wDXdP61z4n+DL0M6vwM8pooor5g8oKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArX1T/kX9E/64y/+jWrIrX1T/kX9E/64y/8Ao1q0h8MvT9UVHZmRRRRWZIUUUUAFFFFABRRRQBNZ2z3t9BaxFQ88ixqW6AscDP51Z1jSJ9E1FrK7eN5FUMTESRz9QKXQP+Rk0z/r7i/9DFa/xB/5G2X/AK5J/Kt1BOi59bmiiuRyKR8PY8IjXPticybPI28/ex1z174x0rFrq20awHw3TVRB/ppkwZd7dPMI6Zx09q2NW0bwnov2GbULeYLMmPKidzuPGWPzZAGe3r3rZ4dySastE9+5Tpt67aI88ort9X8GWp8VWFnprGK1vELtyW2BeTgn1GMe9aq+C9PnuprQ6JNawqpEd/8AawxY+uzPH5flSWDqttdgVGbdjzOius0LStJRLtNVsL7UbqC4aLyrSNyABxnjGOc9TVzXfC+nWesaK9rDJHa31wkctvIxyMsvHqOCQeahYabhzJoXspWucPRXb+JtP8L6I1xaLazG8ki3Q7HYrCccZy3OTn1riKzq0nSlytpkzjyOzCiiisiAooooAKKKKACtfxT/AMjRff8AXT+grIrX8U/8jRff9dP6CtF/Dfqv1K+yyvo+kT63qK2Vo8aSMpYGUkDj6A0axo91oeoNaXoXeFDBkJKsD3BOPp+FbHw+/wCRti/65P8Ayre1aNPFthqFsgH9qaTcSCMDrIgYgD8QPzA9a6adCM6PMvi/yNY01KF+px+peH7rS9Lsr+4khaK8UNGEYlgCoPOQPWsqu/8AEFm+oeHPC1nEcPMscYJ/hyijP4VdXwXp8V5HZnRJp4NuH1A3YUg46hM/0/OqlhJSlaG2n4obotv3TzOiu3XwtY2uj6/9pj864sXYQylmGBsDDgHHeqeh6NYXngnVL+5g33VuXEcm9htwoI4BweTWP1ad7Ps39xHspXscpRXfDRfDVj4V07VdUtpMuqlxFI26ZiDxjOB68Y6VT0ey8O3Mc8x0vUb/AH3LiOKBHIijz8uSCBnHqTT+qyuk2tQ9k72uczpVj/aerW9l5nlee4Tftzt/CpNd0v8AsXWp9P8AO8/ydv7zbtzlQ3TJ9a6670C00Txtokmnh0huZM+U5JKEY9ee/esHxz/yOl9/2z/9FrTqUfZ0nzbp2/AcqfLF33uUNE0S516+a0s3iSRYzITKSBgEDsD61FqmmXOj6jJZ3igSR91ztYdiD3FdF8N/+Rnl/wCvVv8A0Ja0dagTxXotxPAM6npUjRyKBzKgJ5/IZHvkVUaEZ0eZfFr+A1TThdbnK634eu9BW2a8khcXKlk8picYx1yB61lV6H4zsX1W98PWcLBTOrKG6gD5Mn8qsjwXp63i2R0Sd7fZg6ibwA7sddmfX2/CqnhJOo1DZW/IcqLcmonn1jpl5qXm/YrdpRCu+QjACj3J+hqpXqHhe207S4dXsDDI1xagi6kzxKmWK454O36Vhrpug6z4f1i60eye3ltMPFvkYsFCgnI3EckOKl4V8iaeuunoJ0dFZ6nF0V3MnhSxXwK1wIv+Jolst0zl2yFJJxtzj7oI6dqG8JWU40TT44jBd3EP2i7m3ksqgDjaTjknHTtU/VKn5fiT7GRw1FegRWHgu61ZtGit7hbncYxPvbBYdQOevHpiodB8JWDahrllqiiU2ewRS7mXaGDHdgEdgDzR9Um2kmncfsZXsmYPhjw5/wAJHdTw/avs3lIHz5e/POPUVh16d4Ql0STWZx4fgmiVINsrSk/OdwwwyT7+n0rnrjSNN8O+HVk1m3Fzqt0N0MBkZREPVtpH/wCvj1NXLDr2akmut30G6fupo5KtfVv+QLon/XtJ/wCjXrIrX1b/AJAuif8AXtJ/6NeuWHwy9P1RlHZmRRRRWZIUUUUAFFFFABRRRQAUUUUAa+r+Gr3RmtFuGila7B8pYSSe3HIHPIqa+8J3unXNjbXNxai4vZFRIldiy5OMt8vAycd/bNemvYW89zY3si+bcWluxghyBkkDn9MfjXmcF/d6l46s7i/yJjfRKUPHl4cDbjtivRrYenSfq9P1OmdOMH6lfXfDd94ekiW98t1lBKyRElcjqOQOaQ+HrseGxrfmQ/Zi23ZuO/723pjHX3r0HWDBrmo3/hy8ZVmKLPZuR907RkfzP0J9KxruCS1+FBt51KSxTlHU9iJTTnhoKUmtkn96CVJJu2xz9x4R1G30BNXzDJbuiybY2YuqnuRjHHfmq+h+HrvX2nWzkhQwKGbzWIznPTAPpXoCarDpfhvQftgVrS6hSCYMOAGTqfYY59iaj8N6I2ha/q1uMmCSFZIHP8S5PH1HT8vWq+qwc4226/dcfsYuSseXUUUV5ZyhRRRQAUUUUAFFFFABRRRQB3fww/4+9R/3E/ma9Frzr4Yf8feo/wC4n8zXotfRYH+Avn+Z6WH/AIaCiiiu03CuV+In/Iqn/run9a6quV+In/Iqn/run9a58T/Bl6GdX4GeU0UUV8weUFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVr6p/wAi/on/AFxl/wDRrVkVr6p/yL+if9cZf/RrVpD4Zen6oqOzMiiiisyQooooAKKKKACiiigC5pE8drrljcTtsiiuI3dsE4UMCTxXZ6xL4M1vUWvbvV7pJGUKRFGwHH1jNcBRW9Os4RcbJrzNIz5Vax191rGljwA2k210XnWdiiMjZK+aSDnGOmDUfjXWrDV49NGnz+cYI2EnyMu0nb6gehrlKKcsRKUXG3RL7gdRtWO71TxfYxa3o97YSG5S2iaOdQhU4IA43Ac9/wAKbfSeB7q8l1F57l5ZMu1sqMFZjz6ev+1iuGoq3ipO/Mkyvat7o7XSNd04eDo9MfVZ9JuY5CzSxRMxkG4nqvsQOo6elS6t4i0m4/4R8W15LKLK4jeZpkbeFBXJY45PHbNcLUsVrcXEcskEEsqQjdIyISEHqSOnShYmfLypf0g9rK1jY8Y6ja6r4ikurCXzYTGoDbSvIHPBANYVFFc85ucnJ9TKTcndhRRRUCCiiigAooooAK1/FP8AyNF9/wBdP6CsitfxT/yNF9/10/oK0X8N+q/Ur7LJvB2o2uleIo7q/l8qERsC20tyRxwATTU11tP8Y3Gq2Lb4nuZGxyPMjZicfiPXvisSimqslFRXR3HztJJHe+J/FmnTSaVcaPN5z2k/mtGY2XA445H4cUaheeC9bul1G9uLmG4ZR5kKow34GMHCke3BFcFRWzxc5NuSTuW6zb1R2XhvXtGtotV07UA8NhdyM0R2scKeNpxkg4x+tWhrXhux8K6lpel3EuZFYIZUYtMxUc5AwPTnHSuDopRxU4xtZf8ADiVVpWOr1zWrC88E6XYW0++6tyhkj2MNuFIPJGDyau2ev6ZL4StNP/tWfSJ7c/vPIiYmQc5wV6Zznr19a4eiksTNS5rdLB7V3ud5qviTSbnXNAuLe6ZobQkzM6NuT7vXjk8Hpmub8V31tqXie7u7KTzYJNm19pGcIoPB56g1j0UqmIlUTT6u/wCFhSqOSszovBOq2eka69xqM3kxGBkDbS3JKnGAD6GotN186T4rlv4GL28sz+Yo/jjLZ6Hv3rCoqVWlGKS6O4lNpJLodz4l8VWD6vo95o0ouBZFiybGQYO0Y5A7A0/ULrwVq95/aV3c3MczAGSFUYbzjHPyn9CK4OitXipttySdy/bNt3R1vhLXNK02+1OO93wWl4MIdpO1cn5TjJ6N+lO0HVdH0PxNexQ3Uj6RPCEEsiEknAOSAuepYdO9crFa3FxHLJBBLKkI3SMiEhB6kjp0qKpjiJxUdNthKo0l5Hdw+LLBvGl5LcTY0ua2+zqfLbGAOOMZ6lu3eql34vii8djU7bM1nHGIAFXBaPHOAcdyTXH0U3iqjVvO/wDXkHtZfqd/Hf8Agu01ZtZguLh7nLSC3CNgOc5xkYzz64qto3iqzafxBdanKLeS/VRCgVmzhWAGQPQjriuJop/WppppJf1YftnfRHU+BdZsNG1C6k1KfyUkiCqdjNk59gam/tvTNf8ADIs9fuPs+oWvFvcmNn3j3wD6YP4GuQoqI4iUYKFtNfxJVRpcvQK19W/5Auif9e0n/o16yK19W/5Auif9e0n/AKNes4fDL0/VEx2ZkUUUVmSFFFFABRRRQAUUUUAFFFFAHca/4tt1vtIu9FufNe1RlmXaygg7flORznBqPWNS0C71/TNasrry5RcRPdQtE4IAIJbpjIAwcde1cXRXVLFTle6WtvwNXVk9zo/E2txXHi0ano9wWEaoUkCleQPQgVt+I/Fel6x4RaGGTZeylGeDY3DAjPzYwelcDRS+sz97+8HtZa+Z1fiDWrC98H6TY2s++4twglTYw24TB5IwefStfwx40sYtGFtrU5jngUxxSFGbeh6DgHpgD8q89opxxVSM+ddrAqslLmCipWtbhLVLl4JVgc7UlKEKx9Aeh6Goq5rNGQUUUUgCiiigAooooAKKKKAO7+GH/H3qP+4n8zXotedfDD/j71H/AHE/ma9Fr6LA/wABfP8AM9LD/wANBRRRXabhXK/ET/kVT/13T+tdVXK/ET/kVT/13T+tc+J/gy9DOr8DPKaKKK+YPKCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK19U/5F/RP+uMv/AKNasitfVP8AkX9E/wCuMv8A6NatIfDL0/VFR2ZkUUUVmSFFFFABXp13r11oHgvRJrKKKV5Y40KyKTxszxgjmvMa9QuPEE3h7wbodxDCkyyRxo6vkcbM8Hsfzruwjspu9tN/mb0Xa+tjL8eW8d1pulXv2TyNQucBolHzHKg7SOpIJA/GshfAupnbG1xYpdMm8WjT/vcfTGP1rpNShhm17SPE0dw02nO6q+85EBPAPsM9fQir+p22uJ4g+0aVpGkTIxBW7ljxIvGPmO4H8h0rqnRjUm5yXbb8/maypqUm3/XmefaT4Z1LWbya3to1ja3O2VpTtCHOMHrzwa0LvwFq1laXVzPJaiO2QuSHY7wBnj5f54rpNIku9X0PxBbwSww6m9027yiQOijjJyAdrDNN0vSb/R/AesQakwV2hkZId4byxtPpxyayjhqbitG9G7kqlG2xzNl4H1S/sLW7tntzHcnjLsCgweW46cY4z1FRQ+D9RluLtHktoIrRyklzNIVjz7HH9K3dYuJYPhXpixOVEsio+D95cOcfmBV/QPMvvh7FDp0Fne3UMjeZBdruUncT0yOcEYJpKhSclGzva/8AwAVODdvK5xt/4Y1KwuLaLZHci7IEEtu+5JD6A8VauvBl7ZwTvNe6f5sERle3WcmQKBn7uK3b668R2s2j28um6fA/ntLb21sNv3RyD8xUAhz0qzqekWXiKwvdTubG80m+gjJdphhXIX36jjGRij6vB83Knfz06C9nHWxy1n4Pv7rT4rya4s7KGYjyjdTbC+emOD1rS0e01zS9N12GzlsvKt1YXBfcTwpJKYHp61r6bH/bmj6dpWvaPdBFiU295D9wLt4JI6HGODnmk0ywj0vRvFdlDKZkhjdVc9f9UTg+46fhVwoxi1KP39dio00rNf1ocponhLU9dgae1WOKEHAkmYqGPtgEmo9b8L6joCo96qPE5wJYmyufTkAiurazudf+HNhb6GwLwsFnhDhN2Acg9upB5+tR6zFLo/w3j0/VpVN5JIPLj3bioDZxn2H4c4rN4eCp3s9r36ehLpx5fluZf/CutZ+1eSHtSNm4yb22jnp93OfwrB1bSbrRdQazvVUSABgVOQwPcV2HxIu7i31XTvJmdPLQyLtPRs9frVL4l/8AIyW//Xov/ob1NejSipcu8WvxFUhFJ26HH0UUVwHOFFFFABWv4p/5Gi+/66f0FZFa/in/AJGi+/66f0FaL+G/VfqV9lmRRRRWZIUUUUAenXevXWgeC9EmsoopXljjQrIpPGzPGCOazfHlvHdabpV79k8jULnAaJR8xyoO0jqSCQPxrUuPEE3h7wbodxDCkyyRxo6vkcbM8HsfzqDUoYZte0jxNHcNNpzuqvvORATwD7DPX0Ir26nvQcL30WnbzO6Wqtft8vM5tfAupnbG1xYpdMm8WjT/AL3H0xj9ao6T4Z1LWbya3to1ja3O2VpTtCHOMHrzwa9B1O21xPEH2jStI0iZGIK3cseJF4x8x3A/kOlUtIku9X0PxBbwSww6m9027yiQOijjJyAdrDNYPC0+fls+vzIdKPNY5u78BatZWl1czyWojtkLkh2O8AZ4+X+eKbaeBNVvbO1uYJLYx3Khhl2BQYzlvl/DjPWuk0vSb/R/AesQakwV2hkZId4byxtPpxyapeIZpIvhnpCRuyrJ5auAfvDYTg/iBSdClFczTWl7fMTpxSu10MC28JahcNdl5La2gtZWie4uJNkZZTg4OOfyqHUPDOo6fd21uUS4N3xBJA25ZPofxrtdJ8zUPh/arpVrZX08LnzILxAw3ZbPfhjnOT2NUb+78R293o1vNp1hFKsrSW9tbfL90YKn5ioBDHpSlh6agnrrb8ROnHluYd14MvbOCd5r3T/NgiMr26zkyBQM/dxTLPwff3WnxXk1xZ2UMxHlG6m2F89McHrXU6npFl4isL3U7mxvNJvoIyXaYYVyF9+o4xkYo02P+3NH07Ste0e6CLEpt7yH7gXbwSR0OMcHPNP6tDntb08/XTQfso3MjR7TXNL03XYbOWy8q3VhcF9xPCkkpgenrWXovhDVNdtjcWyxRQZwskzEBj3xgE11WmWEel6N4rsoZTMkMbqrnr/qicH3HT8KZc2V14g+H+mRaG4cwhVmhDhNxAwQc8deefrT9imle7snp8w9mmlf7vmcpq3hbUtFnhS7WMpMwRJo2JQt6dMj8q0v+Fd6ybwwb7XAQP5u9tnJIx93OePTuK2tSjk0rwXpunarKr3zXKFF3bioD56+w4/HFV/HV3cReMdMSOZ1WNI5FAPRi7An8gKmVClBOUk+mna4OnCOr8jjtU0u50fUHs71QJEwcqchgehBqnXXfEj/AJGeL/r1X/0Jq5GuKtBQqOK6GE48smkFFFFZEBWvq3/IF0T/AK9pP/Rr1kVr6t/yBdE/69pP/Rr1pD4Zen6oqOzMiiiisyQooooAKt6T/wAhqy/6+I//AEIVUq3pP/Iasv8Ar4j/APQhVQ+JDW56drHiS+sPFtppVvaR3EE6IX+U7xliCc5xgAZ6Vyvi7RPtHjQWej248yeJXdE4AYk5J9OAD/8Arrq9X8Wf2P4pg0+6RBZzRKzS5IZCSRn0xwKq6Tpv9jeO7o3MplF/CzW00jZLcglM9yOPwFe1Viqr5G76/d/w53TXO+W/X7jj7rwbf29rPPFcWd39mGZo7abe8frkYFGj+C9V1qyF3b+TFCxOwzMRvx6YBrqZB4i0mO+lj0fRbaCOJmeaKPb5qgZ4AbP51HqWmXuv+BtHTRXWRY4kE0QcLuIUDuccEHiub6vTu9HotjL2cez9Dl9T8IajpNrbz3jwATyiJUViWVjnrxjHHbNXJvh9rEHnF5LUpDF5pcO2D14Hy9ePpyK3vE0E9r4W0OC7l86aO6jV33ZyQG79/Ss34l3M39uWsHmMIltw4UHjcWYE/kBSqUaVNSbT0t+IpU4RTduxk2ng6/ubSG4lns7MXH+pS6m2NJ6YGD1qGHwrqs+tSaWIAs8Y3OzNhAvZs+hr0HXEvb6Ozu9D0zS9ThZBg3MYZl7ggkgY9qyYdX12DxXePc6dFdtDbxxXMVnngH5gQCSSRuIwP/r1UsNSi0nf17jdKCdnc5XUPDFzYQRypdWd4JJhAFtJS5DnseB6Vak8D6hCh8+7sI5xEZTbNP8AvMAZ6AVr6voFrZR6dq+iQTWlzLdRiOyueMvuyBg8jkevStK70q18Uxz3Oo6deaTfQRjdM/CHGeh6MB68fWpWGjdq2vTXT77fmHsldqxzepHWI/ANgLl7Q6e8gEYQN5v8RAbjGOD09qSx+H+tXtolx+4tw4yEmchsfQA4rV1GSOH4d6BJMN0aXcbOMZyAHzVrxboWr65qlpeaPIJbXyl2MswURnJO78iORzVOjGXvNOVktAcE9d9jkR4X1EeIY9HlWOG5kBKM7HYwAJzkA8cGtGP4d61Isp3Wy+WxUBnI347jjp9cV1V7cRN4/wBBtfNWW4t4ZfOI9TGcZ/InHvWRY3dwfi5KhmcqzyRkE8FQhIH0yAaX1ejF2d371vwQezgnZ97HCyxPDM8UqlXRirKexHUUytDX/wDkZNT/AOvuX/0M1n150laTRzNWdgoooqRBRRRQB3fww/4+9R/3E/ma9Frzr4Yf8feo/wC4n8zXotfRYH+Avn+Z6WH/AIaCiiiu03CuV+In/Iqn/run9a6qs7W00yTTtutKGtt44Ib73b7vNY11zUpK9tCKivFo8Qor037L4G/54p/3zLR9l8Df88U/75lrwfq8f+fkfvOD2P8AeR5lRXpv2XwN/wA8U/75lo+y+Bv+eKf98y0fV4/8/I/eHsf7yPMqK9N+y+Bv+eKf98y0fZfA3/PFP++ZaPq8f+fkfvD2P95HmVFem/ZfA3/PFP8AvmWj7L4G/wCeKf8AfMtH1eP/AD8j94ex/vI8yor037L4G/54p/3zLR9l8Df88U/75lo+rx/5+R+8PY/3keZUV6b9l8Df88U/75lo+y+Bv+eKf98y0fV4/wDPyP3h7H+8jzKivTfsvgb/AJ4p/wB8y0fZfA3/ADxT/vmWj6vH/n5H7w9j/eR5lRXpv2XwN/zxT/vmWj7L4G/54p/3zLR9Xj/z8j94ex/vI8yor037L4G/54p/3zLR9l8Df88U/wC+ZaPq8f8An5H7w9j/AHkeZUV6b9l8Df8APFP++ZaPsvgb/nin/fMtH1eP/PyP3h7H+8jzKivTfsvgb/nin/fMtH2XwN/zxT/vmWj6vH/n5H7w9j/eR5lRXpv2XwN/zxT/AL5lo+y+Bv8Anin/AHzLR9Xj/wA/I/eHsf7yPMq19U/5F/RP+uMv/o1q7b7L4G/54p/3zLWj/ZPhi7sbb/Rke3RWEP3xgbjn36561rTwvNdRnF6dH5oqNHdJo8hor1r/AIR3wr/z5J+cn+NH/CO+Ff8AnyT85P8AGn/Z9T+ZB9Xl3R5LRXrX/CO+Ff8AnyT85P8AGj/hHfCv/Pkn5yf40f2fU/mQfV5d0eS1PLfXc9vHBPdTSQx/cjeQlU7cA8CvU/8AhHfCv/Pkn5yf40f8I74V/wCfJPzk/wAaf1CovtIPq8u6PLY9QvYrVrWK7nS3bO6FZSEOeuRnFSLrGpJCIU1G7WIDAQTsFA9MZr07/hHfCv8Az5J+cn+NH/CO+Ff+fJPzk/xp/Uaq+0vvY/YT7nCeFrjSYHuf7Wu7qykYL5NxbMysvXcPlz1+Xt2rY1jxNpVroM+m6HNcXkt3xNc3BYkjGCSWwScDHTFdH/wjvhX/AJ8k/OT/ABo/4R3wr/z5J+cn+NbRw9aEOROPr1LVOajZNHlkl7dS2qW0tzM9vGcpE0hKKfYdB1P50W17dWTl7O5mt2IwWikKk/lXqf8AwjvhX/nyT85P8aP+Ed8K/wDPkn5yf41j9Rq3vzIj6vPueXPqN7JcrcSXlw86fdlaViy/Q5zTrjVdQvI/Lu765nT+7LMzD8ia9P8A+Ed8K/8APkn5yf40f8I74V/58k/OT/Gj6jV/mX3sPYT7nmSazqkcKwx6ldrEo2qizsFA9MZqGK+u7eGWKC6mijmGJESQhX7cgda9T/4R3wr/AM+SfnJ/jR/wjvhX/nyT85P8aPqNX+Zfew9hPueWWt7dWTFrO5mt2bgmKQqT+VNuLq4u5fNu55J5MY3yuWOPqa9V/wCEd8K/8+SfnJ/jR/wjvhX/AJ8k/OT/ABpfUKtrcyD6vLa55ZdXt1esrXtzNcMowplkLkD8aS5vLm9kEl5cS3DgbQ0rliB6ZP1r1T/hHfCv/Pkn5yf40f8ACO+Ff+fJPzk/xo+o1X9pB9Xn3PJaK9a/4R3wr/z5J+cn+NH/AAjvhX/nyT85P8aX9n1P5kL6vLujyWivWv8AhHfCv/Pkn5yf40f8I74V/wCfJPzk/wAaP7PqfzIPq8u6PJa1/FP/ACNF9/10/oK9D/4R3wr/AM+SfnJ/jU95ofhy4vJJby0R5mOXYl+T+FWsDPlauun6j+ryta6PHqK9a/4R3wr/AM+SfnJ/jR/wjvhX/nyT85P8aj+z6n8yF9Xl3R5LRXrX/CO+Ff8AnyT85P8AGj/hHfCv/Pkn5yf40f2fU/mQfV5d0eWS313PbxwT3U0kMf3I3kJVO3APApY9QvYrVrWK7nS3bO6FZSEOeuRnFepf8I74V/58k/OT/Gj/AIR3wr/z5J+cn+NV9Rq/zIf1efc8xXWNSSEQpqN2sQGAgnYKB6YzWn4WuNJge5/ta7urKRgvk3FszKy9dw+XPX5e3au7/wCEd8K/8+SfnJ/jR/wjvhX/AJ8k/OT/ABqo4OrGSlzJ273GqM073RzmseJtKtdBn03Q5ri8lu+Jrm4LEkYwSS2CTgY6Yrj5b67mtkt5rqaSCP7kTyEqv0HQV6n/AMI74V/58k/OT/Gj/hHfCv8Az5J+cn+NVUwtao7uSHKlOT3R5Za3t1ZMWs7ma3ZuCYpCpP5Ur6heSXS3Ml3O06/dlaUlh9DnNepf8I74V/58k/OT/Gj/AIR3wr/z5J+cn+NZ/UatrcyJ+rz7nmFxquoXkfl3d9czp/dlmZh+RNOTWdUjhWGPUrtYlG1UWdgoHpjNem/8I74V/wCfJPzk/wAaP+Ed8K/8+SfnJ/jR9SrXvzr72HsJ9zyyK+u7eGWKC6mijmGJESQhX7cgdaW1vruyLGyuprct1MUhTP5V6l/wjvhX/nyT85P8aP8AhHfCv/Pkn5yf40fUav8AMg+rz7nlc13cXE4muJ5ZZR0kdyzfmaW4vbq7mWa6uZppVACvJIWYAHPBNep/8I74V/58k/OT/Gj/AIR3wr/z5J+cn+NL6hV/mQfV59zyu5u7m9lEl5cS3EgG0PK5Y49Mmoa9a/4R3wr/AM+SfnJ/jR/wjvhX/nyT85P8aHgKj3khfV5d0eS0V61/wjvhX/nyT85P8aP+Ed8K/wDPkn5yf40v7PqfzIPq8u6PJa19W/5Auif9e0n/AKNevQ/+Ed8K/wDPkn5yf41PNofhyS3t0mtEMcaERDL8DcSf1Jq44GaTV1r/AJjWHkk9UePUV61/wjvhX/nyT85P8aP+Ed8K/wDPkn5yf41H9n1P5kL6vLujyWivWv8AhHfCv/Pkn5yf40f8I74V/wCfJPzk/wAaP7PqfzIPq8u6PJacjtG6vGxVlOVZTgg+tesf8I74V/58k/OT/Gj/AIR3wr/z5J+cn+NH9n1P5kH1eXdHldzd3N7KJLy4luJANoeVyxx6ZNPk1K+lhiilvLh44SDGjSsQhHTAzxivUf8AhHfCv/Pkn5yf40f8I74V/wCfJPzk/wAar6jV/mQ/q8+55hPquo3UPlXN/dTR/wBySZmH5E10ugXfh+LR4UvNV1DT7lSxmS3kdVl+Y4zgEdMDt0rqv+Ed8K/8+SfnJ/jR/wAI74V/58k/OT/GtIYSrCXNdP1uVGjNO90cb4q8SW2oRWmn6Orx2dngo5yCzAYGO/A7nnmuduby5vZBJeXEtw4G0NK5YgemT9a9U/4R3wr/AM+SfnJ/jR/wjvhX/nyT85P8ampg61STbkhSozk7to8vt9U1C0i8u0vrmCP+5FMyj8gaSHUb63meW3vLiKST77pKys31IPNeo/8ACO+Ff+fJPzk/xo/4R3wr/wA+SfnJ/jU/Uav8y+9i9hPueWzahe3EyTXF3PLLGco7yszKfYk8VJPq2o3URiutQupoz1SSZmB/AmvTv+Ed8K/8+SfnJ/jR/wAI74V/58k/OT/Gj6jV/mX3sPYT7nlkl7dS2qW0tzM9vGcpE0hKKfYdB1P51Jb6rqFpD5VrfXMMf9yOZlH5A16f/wAI74V/58k/OT/Gj/hHfCv/AD5J+cn+NH1Gre/MvvD2E+55VFdXEFz9phnljnyT5qOQ2T1569zTlvbpLz7WtzMtzknzhId+SMH5utep/wDCO+Ff+fJPzk/xo/4R3wr/AM+SfnJ/jS+o1f5kH1eXc8nkkeWRpJXZ3clmZjksT1JNNr1r/hHfCv8Az5J+cn+NH/CO+Ff+fJPzk/xpf2fU/mQvq8u6PJaK9a/4R3wr/wA+SfnJ/jR/wjvhX/nyT85P8aP7PqfzIPq8u6PJaK9a/wCEd8K/8+SfnJ/jR/wjvhX/AJ8k/OT/ABo/s+p/Mg+ry7owPhh/x96j/uJ/M16LWXpGm6RYPKdIgWJnAD4Lcjt1rUr1sNTdKkoM7KUXCFmFFFFdBoFYPjH/AJAP/bVf61vVg+Mf+QD/ANtV/rXLjP8Ad5+hnV+BnBUUUV8WeWFFFFABRRRQAVnaNr2meILWS40e6FzHFKYnIVlKsOoIYA960a8p+Hh/sF9JuScWevG4t5OeFuY5pDGf+BJlfqorop0lOnKXVWt+L/QuMbps9L1TVLPRtNm1DU51t7WEZkkIJxzjoOTye1Qza/pdtocWsXN7HDp8saypPJlQysMrweckHpjNcj8Q5jqrXGlR/Nb6Zp82p3g7FlRhCh/4EC2P9kVFJpl3ceD/AATqlrYjU49LtYZZrHIzKrQBdyg8FlPIFaxoQ5Iyk9/8tPv/AFKUFZNnTaH418PeJLp7bRdSS5nRd5j8t0OPUBgM9R0qnf8AxL8JaZqE9jfat5VxbuY5E+zSttYdRkLg1oaHr+jeI2abTyv2u2GySGaLZPBn+FlPI6fSuO0PxX/YWq+Jbb+wdb1HfrU8nm6fZ+bGuVQbScj5uM49CKcaMZSkuV6dLq/32/QFFNvQ9C07ULXVtOgvtPl822nXfG+0ruHrggEVBqWs2+l3WnwXCSs2oXH2eIoAQrbS2WyRxhT0zVmzuftljBc+TNB50ayeVOu2SPIztYdmGcEetc34x/5DXhX/ALCw/wDRT1z04qVTlfmQkm7HV0VyOoeIdf1DXrvSvB9lYv8AYNq3V5qDuIg5GdihOScdT/kv0fxTfifUtO8S2EdtqWn2/wBqItGLx3EWD8yZ56gjBp+wny3/AA6j5Ha51dFcb4d1vxbrJstSkstIGj3gDBIZ3a4iU9CSflJHcDmoz4j8VavqmpL4ZsdKNppt01tIl7M4mmZepULwoPYt160/q8rtXWnmHI7nbUU1CzIpddrEcrnOD6U6ucgKKKKACiiigArrtN/5ANj/ALr/APoZrka67Tf+QDY/7r/+hmvVyv8AjP0/VG9D4mTUUUV9AdYUUUUAFFFFABRRRQAUUV5xrNv8SrLTb3xAviPTIvsqPcDRlsA0RjUbtpmJ37sDtgZppXGej0VwuveMNUn8L+GJfDqQ2t/4maFYprobo7UPF5hJH8TADAHc1HouoeKfD3jmx8O+KNXg1+DVLaWa3u47VbeSF4sFgyLxtIYYPrRYLHfUUV5tP451O++MmjaPpMgXQPNubW6lCqwurhIGkZQSMgIdoJBGSSOcUJXA9Jori/iD4o8Q6Dp10PDWi/aWhspLubULhwsFuqKxIx1d/l4UeozxXR+Hrqa+8MaXd3T+ZPcWcMsj4A3MyAk4HA5NFgNGiiikIKKKKACiiigAooooAKnvP+PyT61BU95/x+SfWn0H0IKKKKQgooooAKKKKACiiigAorzf4q+OdT0C3jsfC7hb6Jori+n2K4toGkWNQQwI3OzADvgMfeuw8SWev31lFB4a1S30qZpP311Nbeeypg/dQkAnOOtOwzYorhPC+o+J9M8dz+FfFGqW+tq2nf2hBfx2wt3A8wRlHRfl75GPQ/h3dLYAorzafxzqd98ZNG0fSZAugebc2t1KFVhdXCQNIygkZAQ7QSCMkkc4roPEeneMtV1PyvD2u2mg2EcQPn/YxczSyEnI2sdqqBjnrkmnYLHU0VyfgTXNX1D+19K8SGCXUdGuxbyXNuu1J1ZA6tt7HB5HSuspAFFFFAgooooAKKKKACp5v+Pe3/3D/wChGoKnm/497f8A3D/6EaaGQUUUUhBRRRQAUUUUAFFFFABRRVLWI9Sl0mePRJ4Le+YARS3CF0TkZJUdeM4HrigC7RXnMdz408J+LNEt9d8Q23iHT9YuGtWT7AttJbvsZwy7Cdw+XnP/ANet3xJpHjDVNR/4p/xRbaFZJGMAaetzJI+TnJc4A6dKdhnU0Vyvw/1zVda0nUItf8h7/S9Rm0+We3Uqk5jCneB2zuwR6g/SukvLy30+xmvL2ZILeBDJLK5wqKBkk0gJqK89+G/i3W/FHibxN/bCtBZxC0m060dFDQwyrIQWIGSWVVYgk4zjipfir4yv/Dfh+S18NsDrcsL3CttVvs0EYzJKwII/2RnqW9qdnewWO9oqppU8lzo1lPO26SW3jd2xjJKgk8VbpCCiiigAooooAKKKKACiiigDS0f78v0FatZWj/fl+grVreGxpHYKKKKooKwfGP8AyAf+2q/1rerB8Y/8gH/tqv8AWuXGf7vP0M6vwM4KiiivizywooooAKKKKACuB0fw7d33wjj06SGW01GKSae2EqFHjlW4d4zg9M8fga76itYVHBWXdP7r/wCZSk0cHZaXqM3w78RahqNrINZ1u2uJZLcIS6jyisUQGM5Cgcdck1fs9UvfDngnw+G0LUL7bYwpcR2sYMsBWNeChwSeox2IrraKt1ub4lpe/wCg3O+5w2nyX3iTx1p+sw6Fe6PaWUEqTTX0Yilud4AWPZknAI3ZPp2709K8QXHhjVPEFvc+GfEF2LnVprmKazsDJGyFVAwSRn7p6V6LRVe3T0cdLW/UfOuxS0nUf7W0uG9+x3dl5u79xeReXKmGI+ZcnGcZHsRWP4rtbi41fw09vBLKsOph5WRCwjXy3G5sdBkjk10tFYxnyy5kiE7O5w0l1qXg3xLq0x0W+1bTNUmFykmnR+bLFJtCsrJkccAg1PoyapqGv6j4pu9Jmsx9gFpZ2M7ASyqrM5LDopLHAFdlRWrr3Xw67XK59NjzPTkkufFVhP4Y8Mav4ff7SG1NriPyLaSLB3AJuKs2cYIA/wAE8UFr/U7sab4Q1q28QxuyWuqWiiOJ8HCu0oYAggDhgfSvTaKv617ylbbzf49yvaa3sR24lFrELkhpgg8wr0LY5x+NSUUVyGQUUUUgCiiigArrtN/5ANj/ALr/APoZrka67Tf+QDY/7r/+hmvVyv8AjP0/VG9D4mTUUUV9AdYUUUUAFFFFABRRRQAVwNxNffExJrOwZrHwkWaK4vAf32pAHDJF/diOMF+rDIAHWu6uYBdWssDMyLKhQshwwyMZB9a8+tPg7bWFpHa2PjXxlbW8S7Y4YdVCIg9AAmAKasMl+IVpHqdx4b8EQCKzg1SSRhchAXtkt0VgIv7rnIAbsM1n2ug/8K9+ImjyJql5rY17fYO+qyeddQhFLgpJgfJkYZcY5Brq9R8B6XrHhuy0jVri/u2sSGt9QkuT9qRx/H5gx834Y9qr+HvhzpegayNXlv8AVdZ1JEMcV3q12Z3hU9QvAAz06Z/M076DIviZres6R4djh8P6bqN3PfS+RJPp9qZ5LWPHzSBR1bHC5IGTnPGD53P4sstK8SeCIdM8FeK7S20lrsJbT6Zia53wEEoN3ztklmPHc17PpGlf2TFcp9vvr3z7h5917N5hj3Y+RTgYQY4HaotQ8P2upa/pGrzyTLcaQ0rQKjAIxkjKNuBGTweMEc0JoVyl46fzPhn4jfay7tIujtYYI/ctwateE/8AkS9E/wCwfB/6LWrer6bDrWiX2l3TSJBfW8lvI0ZAYK6lSQSCM4Poak06yj0zS7WwgZmitYUhQuQWKqoAzjvxS6AWKKKKQgooooAKKKKACiiigAqe8/4/JPrUFT3n/H5J9afQfQgooopCCiiigAooooAKy/EuqXWi+Gr7UNO0+bUrqCLdDaQKWaVs4AwMnHOTjsDWpRQB88654gNv8NNUtr3wv4rXVtTube51DU77S/KieQTxtt3bjtUAbEX3Hqa9lZP+E68HId2u+HPtRzgH7JeQ7HIweu3O38VPvV3xJ4ftfFGgT6RqEk0dvO0bM0DAOCkiuMEgjqo7dKreKPCVr4pithPf6lp01qzGK5026MEi7sZGeQRwOo7VV0x3OT8PabL4F+I0GiLfyavBrltLObm8Aa8haLGN8owXQ7iBkcHOK1/iZres6R4djh8P6bqN3PfS+RJPp9qZ5LWPHzSBR1bHC5IGTnPGDe8M+BdK8L3c17BNe6hqM6COXUNSuDPOyA527j0GecADP4CtXSNK/smK5T7ffXvn3Dz7r2bzDHux8inAwgxwO1K+oHjE/iyy0rxJ4Ih0zwV4rtLbSWuwltPpmJrnfAQSg3fO2SWY8dzXpmu+Jr9tTj0DwtZrPq8sImmluciGwjbgPJjkscHCDk4PIArW1Dw/a6lr+kavPJMtxpDStAqMAjGSMo24EZPB4wRzXO6x8MbXVfEt9rcPiXxJpVzfiMTppt+IUIjTavGwnpk8k8sfWndMehueFvDcXhrS5IBcSXl3czNc3l5LgNcTN95sDgDgAAdAB9a2qwPC/hT/AIRj7V/xP9c1j7Ts/wCQteef5W3d9zgYzu59cD0rfqWIKKKKBBRRRQAUUUUAFTzf8e9v/uH/ANCNQVPN/wAe9v8A7h/9CNNDIKKKKQgooooAKKKKACiiigAqK7t/tdlPb+bLB50bR+bC210yMblPYjqD61LWdr2h23iLR5dOvZLiKOQqwktpTHIjKcgqw6HIoA88vfDs3w78RaNrUGt32ui8votNki1qQXE6rKdu6GTAKkYyV6EZ9K7HxX4OHiryM6/rek+SrLt0y78lZc4++MHdjH6mqWifDXStI1aHVLvUdX1y9ts/ZpdXvDceRnglBgAH3xT9e+HlprusSakmua9pU0yqsy6ZqBhSUAADcuD2AHGKq4yr8N5prMaz4Zm+zzDQbpYY7q3hEYnV0EmWUcbwSQ2OprG+Ius3k3im00W98Na/qPh6BVubo6XYGcXkmcpExyAEBGWGSScDAHXstP8AB+maP4Xn0PRmubCK4DmS5gmIuC7dZPMOTv8Af6elbFnbfY7GC286afyY1j82dt0kmBjcx7scZJ9aV9bgeQ+HPHjj4j+LryHwn4kL3VrbOls1hteIwwykLIMnZv6J1yaxL7xNcjwV4nudc8JeK/7c1m0kS4vZtKKW1qm0hI1YtlY1zycZJyT7e12Ph+10/wAR6rrUMkzXOqrAs6OwKKIlYLtGMjIY5yT+FTa5pEGv6De6TePIkF7C0MjREBgrDBwSCM/hTuh3M3wPrH9t+ErO4/s7UNP8uNYfLv4PKd9qr84GTlTng966CorS2Szs4baIsUhjWNS3UgDAz+VS1JIUUUUAFFFFABRRRQAUUUUAaWj/AH5foK1aytH+/L9BWrW8NjSOwUUUVRQVg+Mf+QD/ANtV/rW9UVxawXcXlXMSypnO1hkZrGvTdWlKC6omceaLR5TRXpf9g6X/AM+MP/fNH9g6X/z4w/8AfNeD/ZFX+ZHH9Wl3PNKK9L/sHS/+fGH/AL5o/sHS/wDnxh/75o/sir/Mg+rS7nmlFel/2Dpf/PjD/wB80f2Dpf8Az4w/980f2RV/mQfVpdzzSivS/wCwdL/58Yf++aP7B0v/AJ8Yf++aP7Iq/wAyD6tLueaUV6X/AGDpf/PjD/3zR/YOl/8APjD/AN80f2RV/mQfVpdzzSivS/7B0v8A58Yf++aP7B0v/nxh/wC+aP7Iq/zIPq0u55pRXpf9g6X/AM+MP/fNH9g6X/z4w/8AfNH9kVf5kH1aXc80or0v+wdL/wCfGH/vmj+wdL/58Yf++aP7Iq/zIPq0u55pRXpf9g6X/wA+MP8A3zR/YOl/8+MP/fNH9kVf5kH1aXc80or0v+wdL/58Yf8Avmj+wdL/AOfGH/vmj+yKv8yD6tLueaUV6X/YOl/8+MP/AHzR/YOl/wDPjD/3zR/ZFX+ZB9Wl3PNKK9L/ALB0v/nxh/75o/sHS/8Anxh/75o/sir/ADIPq0u55pXXab/yAbH/AHX/APQzW7/YOl/8+MP/AHzVhLC1jiWJIEVEztUDgc5rsweXzw83KTW1jSnRcHdswKK6D7Hb/wDPFPyo+x2//PFPyr0/Zs35Wc/RXQfY7f8A54p+VH2O3/54p+VHs2HKzn6K6D7Hb/8APFPyo+x2/wDzxT8qPZsOVnP0V0H2O3/54p+VH2O3/wCeKflR7Nhys5+iug+x2/8AzxT8qPsdv/zxT8qPZsOVnP0V0H2O3/54p+VH2O3/AOeKflR7Nhys5+iug+x2/wDzxT8qPsdv/wA8U/Kj2bDlZz9FdB9jt/8Anin5UfY7f/nin5UezYcrOforoPsdv/zxT8qPsdv/AM8U/Kj2bDlZz9FdB9jt/wDnin5UfY7f/nin5UezYcrOforoPsdv/wA8U/Kj7Hb/APPFPyo9mw5Wc/RXQfY7f/nin5UfY7f/AJ4p+VHs2HKzn6nvP+PyT61s/Y7f/nin5U5rWB2LNEpJ6kijkdg5TnaK6D7Hb/8APFPyo+x2/wDzxT8qPZsOVnP0V0H2O3/54p+VH2O3/wCeKflR7Nhys5+iug+x2/8AzxT8qPsdv/zxT8qPZsOVnP0V0H2O3/54p+VH2O3/AOeKflR7Nhys5+iug+x2/wDzxT8qPsdv/wA8U/Kj2bDlZz9FdB9jt/8Anin5UfY7f/nin5UezYcrOforoPsdv/zxT8qPsdv/AM8U/Kj2bDlZz9FdB9jt/wDnin5UfY7f/nin5UezYcrOforoPsdv/wA8U/Kj7Hb/APPFPyo9mw5Wc/RXQfY7f/nin5UfY7f/AJ4p+VHs2HKzn6K6D7Hb/wDPFPyo+x2//PFPyo9mw5Wc/U83/Hvb/wC4f/QjWz9jt/8Anin5U42sBABiUheBx0o5GHKc7RXQfY7f/nin5UfY7f8A54p+VHs2HKzn6K6D7Hb/APPFPyo+x2//ADxT8qPZsOVnP0V0H2O3/wCeKflR9jt/+eKflR7Nhys5+iug+x2//PFPyo+x2/8AzxT8qPZsOVnP0V0H2O3/AOeKflR9jt/+eKflR7Nhys5+iug+x2//ADxT8qPsdv8A88U/Kj2bDlZz9FdB9jt/+eKflR9jt/8Anin5UezYcrOforoPsdv/AM8U/Kj7Hb/88U/Kj2bDlZz9FdB9jt/+eKflR9jt/wDnin5UezYcrOforoPsdv8A88U/Kj7Hb/8APFPyo9mw5Wc/RXQfY7f/AJ4p+VH2O3/54p+VHs2HKzn6K6D7Hb/88U/Kj7Hb/wDPFPyo9mw5Wc/RXQfY7f8A54p+VH2O3/54p+VHs2HKyjo/35foK1aZHBFDnykC564FPrSKsrFJWQUUUUxhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAm7nGefSk3c9QPrXEa3Fp9z44s4NMcJq0cqT3U/nkeVEP4CM4+b0xUXji1EOqxeIZPsd5a2Nq+6zlndXf5gdyBTgnjHPrQB3wOaKpTyXcmnh9MWHzmUFBcEhQMdDjmuT8K6Nd6z4dgv9S13VhdSvL5ggvCqAiRhhRjgcUAdzRXPf8IkP+g9rn/gaf8KP+ESH/Qe1z/wNP+FAHQ0Vz3/CJD/oPa5/4Gn/AAo/4RIf9B7XP/A0/wCFAHQ0Vz3/AAiQ/wCg9rn/AIGn/Cj/AIRIf9B7XP8AwNP+FAHQ0Vz3/CJD/oPa5/4Gn/Cj/hEh/wBB7XP/AANP+FAHQ0Vz3/CJD/oPa5/4Gn/Cj/hEh/0Htc/8DT/hQB0NFc9/wiQ/6D2uf+Bp/wAKP+ESH/Qe1z/wNP8AhQB0NFc9/wAIkP8AoPa5/wCBp/wo/wCESH/Qe1z/AMDT/hQB0NFc9/wiQ/6D2uf+Bp/wo/4RIf8AQe1z/wADT/hQB0NFc9/wiQ/6D2uf+Bp/wo/4RIf9B7XP/A0/4UAdDRXPf8IkP+g9rn/gaf8ACj/hEh/0Htc/8DT/AIUAdDRXPf8ACJD/AKD2uf8Agaf8KP8AhEh/0Htc/wDA0/4UAdDRXPf8IkP+g9rn/gaf8KP+ESH/AEHtc/8AA0/4UAdDRXPf8IkP+g9rn/gaf8KP+ESH/Qe1z/wNP+FAHQ0Vz3/CJD/oPa5/4Gn/AAo/4RIf9B7XP/A0/wCFAHQ0Vz3/AAiQ/wCg9rn/AIGn/Cj/AIRIf9B7XP8AwNP+FAHQ0Vz3/CJD/oPa5/4Gn/Cj/hEh/wBB7XP/AANP+FAHQ0Vz3/CJD/oPa5/4Gn/Cj/hEh/0Htc/8DT/hQB0NFc9/wiQ/6D2uf+Bp/wAKP+ESH/Qe1z/wNP8AhQB0NFc9/wAIkP8AoPa5/wCBp/wo/wCESH/Qe1z/AMDT/hQB0NFc9/wiQ/6D2uf+Bp/wo/4RIf8AQe1z/wADT/hQB0NFc9/wiY/6D2uf+Bx/wo/4RMf9B7XP/A4/4UAdDRXPf8IkP+g9rn/gaf8ACq1jaTaV42itBqeoXcEtg8hjupzIAwcDI444NAHVUUDpRQAUUUUAFFFFABRRRQAUVU1O7nsrB57W0N3IuP3QlWPI7ncxAGK5mz+II1HUILOx0iWSaRSZQ1zEojw5U4JPzcg/dzQB2NFc5P4wig1w2X2GZrZJ0tpLwMuxJX6Ljqfr0Fbl7dxWFhPeXJIhgjaRyBkhQMnigCeiucj8a2ksayR6ZrLI4DKRp8nIP4U7/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhornv+Extv+gVrP8A4L5P8KP+Extv+gVrP/gvk/woA6Giue/4TG2/6BWs/wDgvk/wo/4TG2/6BWs/+C+T/CgDoaK57/hMbb/oFaz/AOC+T/Cj/hMbb/oFaz/4L5P8KAOhorK0jxDaazcXMFvFdQzWwUyR3MDREBs4Iz16GtWgAooooAzpdB0ma+N7Lpdo90SMztApc46c4zTptD0u4kjefTrWVov9WzwqxT6ZHFX6KAEAxXPeBP8AkT7b/rrN/wCjXroq53wJ/wAifbf9dZv/AEa9AHRUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGV4i1J9H8PXuoRJveCIsqnua858AfEDWtX8Vpp+pyLPDchiuEA8vAJ4wOnGOa9UuooZ7aSK6VWhdSHDdCK5jwtovhKz1K4m8PPDLdKSH2y7zH7Adq6qMqapSUo3fc460Kkq0XGVkdcKwJv+SiWv/YNk/wDRi1v1gTf8lEtf+wbJ/wCjFrlOw36KKKACiiigAooooAKKKKAKOrrcyadJHZ2dveM/ytBcPtR1PUHg/wAq4lvBWpf2T9iXTdMEkjM0dwJCGsQXztj+XJA69Rya9EooA4p/Cmqf2ibZZIJNOlvYr2WdnPm7kxuXbjByQOc1qeN9Mt9R8I6j9pMw8m1ldfKmePnaeu0jI9jXQ1k+Kv8AkUNW/wCvOX/0A0AW9K/5A9l/17p/6CKtM6opZ2CqOpJwBVXSv+QPZf8AXun/AKCKz/GRI8GaoQSD9nbkVMnyxbNaNP2tWNPu0vvNsEEAg5B6EUVT0g50WyJ5/cJ/6CKuU07q5E48snHsFFFFMkKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMDTv+R61r/r3tv5PW/WBp3/I9a1/17238nrfoAKKKKACiiigArnfAn/In23/XWb/0a9dFXO+BP+RPtv8ArrN/6NegDoqKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiqWs6h/ZOhX+peV532O2kn8vdt37FLYzg4zjrivObT4ua7qFqlzYfD/Ubq3kzslhld0bBwcEQ4PIIrGpWhTdpP8AM9HC5ZisZB1KMU0na7cVr82jtfGjFPBeqMHZCID8y9RXj3wsYr4/tFDsu5JMhejfKetb/iP4jeIr3w7e20/gPU7GOWIq1xIZNsY9TmIfzrgvBmv6ho/ii2u7HRrjUZVDAWsO7dJlT6KTx16dq7sNjsPGhOLbu/7r/wAjmxXDuZSxFOUYxsv+nlP/AOSPp2sCb/kolr/2DZP/AEYtcj/wtDxP/wBE21b85f8A4zU/gnxj/wAJt4nj1L7D9h8q1mg8vzvMzhomznaP73THauCGIpzlyxevozsxGU4zDUnWqRXKraqUXvts2ejUUUVueWFFFFABRRRQAUUUUAFFFFABWT4q/wCRQ1b/AK85f/QDWtWT4q/5FDVv+vOX/wBANAFvSv8AkD2X/Xun/oIrJ8e3AtfAerzMCwW3PArW0r/kD2X/AF7p/wCgisL4k8/DfWv+vY/zFbUKcatWNOWzaT+ZdObpzU47rU2dBk87w7p8gGA1tGf/AB0VfrL8MHPhTS/+vSP/ANBFalZzioycV0FJtybYUUUVJIUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYGnf8j1rX/Xvbfyet+sDTv+R61r/r3tv5PW/QAUUUGgDEvfEY07VIra7sLpIJZVhW8wvlb26Drn8cVHrfiy10S68iW3uLgpEZ52hUEQR5xubJHH09KztStddvfGUD3Glx3Gj2ZWS323QUmTu7qR27CofEGia1eXl1Np9pCf7VsfsdwJJgPs5BOHHHzDDHigDs4pFkiWRWBVgCD6g1geBP+RPtv+us3/o16030m1udJj0++gS4gVFUpIMg4FZXw/iSDwTZxQqEjR5lVR0AErYFAHSUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAY3jL/kRde/7Btx/6KasX4R/8ks0j/tt/6Petrxl/yIuvf9g24/8ARTVi/CP/AJJZpH/bb/0e9cz/AN4X+F/mj24/8iaX/X2P/pMjR8eDPgTVgO8B/mK8d+F4z8Q9Px0Ak/D5Gr2rxbdNZeEtSuEjWQpASEYZBrx/4U3rxeOoohChFyjhm28p8pbj06Yr3sJf6tUsj4zFpfWqd2e9CvE/gb/x9N9Lj+UFe1ivFPgb/wAfTfS4/lBXg1P48PR/ofZ4T/kV4n1p/wDtx7ZRRRXSeIFFFFABRRRQAUUUUAUNas1vtJmgKCRmGUQztEGYcgFl5AzXk1ld3c8OrxX9/PFLo9kZrWNLhiqTebJ8ob/loPlQc169qOmWWrWhtdSto7mAkExyDIyKpt4V0J47WNtJtSlmc24MY/dnOePxoA4eC8up72PVZ7mZdRGr29p5PmEL5LbcjZ05yTn2rrfG8l/H4R1H+zreCbNrKJfOmKbV2nkYU5PtxWk2h6W+rrqjWEBv1G1bgoN4H1qDxV/yKGrf9ecv/oBoAt6V/wAgey/690/9BFYnxH5+HOtf9ex/mK29K/5A9l/17p/6CKxfiLz8Otb/AOvVq6cJ/vFP1X5gX/Cxz4R0r/r0j/8AQRWtWR4T58H6T/16R/8AoIrXrKr/ABJerHLcKKKKzEFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGBp3/ACPWtf8AXvbfyet+sDTv+R61r/r3tv5PW/QAUUUUAFFFFABXO+BP+RPtv+us3/o166Kud8Cf8ifbf9dZv/Rr0AdFRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAlLUVxMtvbSzP92NC5+gGa810P4stqniiKwlsVjtbiTy4nDEsCemRWtOjOom4rYxqVoU2lJ7np9FIOlLWRsFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGN4y/wCRF17/ALBtx/6KavP/AIdfEXwtoXgHTtN1XVPIu4fN8yP7PK23MrsOVUjoR3r1aaGK5gkguI0likUo8bqGV1IwQQeoI7Vk/wDCG+GP+hc0n/wBi/8Aia5qlOo6inBra2p7WExeEjhJYXFRk05KXutLZNdU+5x3if4l+ENV8L39lZ60PPmiKpm1mHP/AHxXm3gDX9K0bxdbX+p3wtraFX3Exu27KkDAVT3NeueM/C2gWng3U57TQdMhmSElJI7OMMpz2IFeVfDjTrG98c2VveWdvcwur7opoldT8hPIIPevSwyxvsJ8rjbro/8AM83EzyH6xT5oVb9Peh/8ies/8Lc8Ef8AQb/8lJv/AIiuG+Bv/H030uP5QV6l/wAId4Y/6FzSf/AGL/4mqFtplhpXj22g0uyt7KJtPlcx28SxqW8xBnCgc4A59q8yNOq6inNrS+3merVxmBhhKmHwsZ3m4tuTT+G/ZLudTRRRXUeEFFFFABRRRQAUUUUAFFFFABWT4q/5FDVv+vOX/wBANa1ZPir/AJFDVv8Arzl/9ANAFvSv+QPZf9e6f+gisb4hjPw71r/r1atnSv8AkD2X/Xun/oIrH+IIz8Pda/69Hrowv+8Q9V+YFvwic+DdI/684/8A0EVsVjeDznwVo/8A15x/+gitms638SXqxy3YUUUVmIKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMDTv+R61r/r3tv5PW/WBp3/I9a1/17238nrfoAKKKKACiiigArnfAn/In23/XWb/0a9dFXO+BP+RPtv8ArrN/6NegDoqKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKDRQaAIrmFbm1kgfO2RCp+hGK8z0P4SyaX4oivpr6OS1t5PMiRQd5I6Zr0522IWYgBRkk9hXNaf8QNA1PXDpdvcN5+SqMyYVz6A10UZ1YxkofM5q1OjKUXU36HTjpS0g6Utc50hRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAVryGC5s5obwK0DKRIG6EVy3hTw74U03VJbjQbiO5usEECcOYx7DtWl45JXwPqpVih8g/MD06V478LmI+IViA2AwkyB3+Rq78PRlOjOSlax5+IrRhXhFxvc+gR1rBm/wCSiWv/AGDZP/Ri1v1gTf8AJRLX/sGyf+jFrgPQN+iiigAooooAKKKKACiiigCpqT38doTpUMEtxkACdyqAdySATXHQfEC7udStLIQafbPJvWSS5uWVGZXKkRkD5s474rpvE2n6nqmhzWejXsdjcSkKZnQthM/MBgggkcZrFufC2rXOgJpAk0mG2MXlSGO2csoz1QluD7nPNAEt34vubfWJI0s43063uYrWefzDvDv3VcYIHetTxW6J4Q1Xeyrmzlxk4z8hrHk8F3H9obIb5Bpkk0NxNE6EyM8fo2cYbAzxWj400+zvvCOpfbLaOfyrWV49652naeRQBp6V/wAgey/690/9BFZPj4Z+H+tf9ej/AMq1tK/5A9l/17p/6CKyvHgz4A1r/rzk/lXRhf48PVfmBL4M58EaN/15x/8AoIrbrD8Fc+BtG/684/8A0EVuVFb+LL1Y5bsKKKKyEFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGBp3/I9a1/17238nrfrA07/keta/697b+T1v0AcNpl14tN1cjSIo7nTAx8mXWC0UpOeQAoyV9CQPxrtIDKbeM3AVZSo3hDkA98e1cLe+LfEdnqOo2pttHQ2MYkPn3JQuCMjGevH68V2el3Zv8ASbS7ZdpnhWQgdiRmgDCsHvbf4kX9pJqFzcWslglwkEpG2JvMK/LxnGAKo6lcalpvjN7zU5tSTR3eGK3aCZBCrsSCHTG45JXnpW0fCOmHXpdX/wBLF5LH5bOLpwCuc425xUs/hnTrrUI7y5WeZ42VlR52Me4dDtzjIoAv3txPb2ryWtq93IvSJHClvxbisT4fsz+CbNnQxszzEoSCVPmtxkV0YGK57wJ/yJ9t/wBdZv8A0a9AHRUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUGgCrqKGTS7pE+80LgZ9cGvnfwdayt4606IAb47kFufQ819D6p/yCLz/AK4P/wCgmvnHwgSfGWknnP2tPx5r1sBf2VT0PHzC3tadz6YFLSUteSewFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBheMraa88G6nb26GSV4CFQd/8AOK8g+F2n3bePLWb7PIEt1fzWKkbMoRz+Jr2TxTfzaX4X1C9tcebDCWQn1ryv4a+KtXuvGUVpdXTzw3SuXD84wpOR6dK9PCuf1epy7HlYrk+s0+bc9rFYM3/JRLX/ALBsn/oxa36wJv8Akolr/wBg2T/0YteYeqb9FFFABRRRQAUUUUAFFFFABRRRQAVk+Kv+RQ1b/rzl/wDQDWtWT4q/5FDVv+vOX/0A0AW9K/5A9l/17p/6CKzPHPPgLWv+vOT+Vaelf8gey/690/8AQRWb43GfAmtf9eUn/oJrow38eHqvzE9g8Ec+BNF/68ov/QRW7WD4G/5ELRP+vKL/ANBFb1RW/iy9WXLdhRRRWRIUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYGnf8j1rX/Xvbfyet+sDTv8Aketa/wCve2/k9b9AHmni+8eLx1F/aNmstmsLiDFr5hkby8qS3+/xiu+0lrhtHs2vVCXBhQyqowA2BmuMk0u4Ov8AiRm8UX8KxJHJLGI4xhTH/CSPTjPFdhobRPoNg1v5hiMCbDL97G0Yz70AaFFFFABXO+BP+RPtv+us3/o166Kud8Cf8ifbf9dZv/Rr0AdFRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABQaKD0oAqaoSNIvCvUQPj/vk185+EHceNNKb+I3SA8e9fRmqf8ge8x/zwf/0E185eEBnxnpP/AF9pz+NetgP4VT0PGzD+LTPpelpKWvJPZCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAwfGo3eCdUBXf+4PHTNeO/Cxc+P7PKbsK5yD935DXsfjLy/8AhDdU+0lhH5DbtnWvHfhV5Q8eWvmFw+x/Lx0J2nr+Ga9XCf7tUPIxf+9U7Hv4rAm/5KJa/wDYNk/9GLW+KwJv+SiWv/YNk/8ARi15R65v0UUUAFFFFABRRRQAUVS1iS7i0W8k05Q12sLGEHu2OK5vQdS1+88RJJfZitJ0fNm6qHtwoAVjg5yxyfpigDr3dI13SMqr6scUhmiAUmRAG+6dw5+lY3i+80Ww8OzXPiSOKWzQj91KARI+flXB4yTivMjb2cNpdjU5rd1/stptJRJAyQ3LSSHZGf7w/djigD2cyIJAhdd55C55/KsvxV/yKGrf9ecv/oBrz5CDqiSXpH/CRDUbVYt3+t8nAzjvt+9mu28bx6g/hHUf7Ont4cWspl86IvuXaeBgjB9+aANbSv8AkD2X/Xun/oIrP8ac+Bta/wCvKX/0E1oaV/yB7L/r3T/0EVn+M+fA+tf9eMv/AKAa3w38eHqvzE9iPwJz4A0P/ryi/wDQRW/XP+Av+SfaH/15Rf8AoIroKVf+LL1f5ly3YUUUViSFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGBp3/I9a1/17238nrfrA07/keta/697b+T1v0AeSeJtR0/VNXaS7ZWls0G7zdLdmX3JHUema9Q0yQTaXayoysrxKwKrtB4Hbt9K4XxBNGfG13ZWd1e2f2i3P214tmx9kW4L8wOCU4yMV2uhtC3h+wa0VkgNuhjV+SF2jANAGhRWba6/p17q91pltcb7y0UNNEY2G0E4ByQAefSmXHiTS7XV00y4uwl25ACbGIyegLYwCfQmgDVrnfAn/ACJ9t/11m/8ARr10IOc1z3gT/kT7b/rrN/6NegDoqKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK4w/FvwR/wBBv/yUn/8AiK7OvIvgvoGj6r4Lu59U0mxvZl1B0WS4tkkYL5cZxlgeMk8e9c1WdRTjCFtb7+Vj2cBhsJPDVsRiua0HFJRaXxX7p9jo7z4s+DGsJ1g1kNKY2CKbWbk44/grxnw34httO8WWWo3UyRxrOGlbyyQqk8kADNe+6h4S8Nw6XdSx+HtJDpC7KfsMXBAP+zXgvhaC3n8XaclxaW0qTXKrJFJCrKQTzwRivRwaxvs58jjb0f8AmedjZZD7Snzxq38pQ/8AkT2kfFvwRjnW/wDyUm/+Irb8PeLtE8U/aP7BvftX2bb5v7p027s4+8oz909PSkHg7wx/0Lmk/wDgDF/8TXE/CmGK28Z+O4LeNIoo9QVEjRQqooknAAA6ADtXmKdaNSMZ2s77X7ep7Dw2XVsJWrYdTUoKL95xa1kl0iu56hRRRXWfPhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBR1eW0g0m6k1LH2RYyZQ3TbXn3w+v/AAfJ4ilj0Wynt7xw3ltO2dy99vpXfa3pi6zol3p7sUFxGU3DtXCeCfhpeeH/ABGupahdQusO4RLESd2QRk5HHBrsoun7KalKzOGsqntoOMbo9MrAm/5KJa/9g2T/ANGLW/WBN/yUS1/7Bsn/AKMWuM7jfooooAKKKKACiiigDO8QpDJ4b1BLqc28LW7h5gMlBjrXGeEbeH/hJre5umu49Rlild3uLcxi6UhQu32UAce9dd4jjv59La207T7a/E+Y5orm4MK7COcEK3NYPhbQtZstaW41myj2RxGOGU6m9w0QOPlClB19c54oA7Ce3huY/LuYY5kznbIoYfkajGn2QRFFnAFjO5B5S4U+o44qxRQBEbaBrgXDQRmZRgSFBuA9M9azvFX/ACKGrf8AXnL/AOgGtasnxV/yKGrf9ecv/oBoAt6V/wAgey/690/9BFUPGP8AyJGtf9eM3/oBq/pX/IHsv+vdP/QRVHxh/wAiTrX/AF4zf+gGt8P/ABoeq/MT2K/gHn4eaF/14xf+giuhrnfh/wD8k70H/rxi/wDQa6KliP40/V/mXLdhRRRWJIUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYGnf8j1rX/Xvbfyet+sDTv+R61r/r3tv5PW/QB5traW+seNrpToEGui3RDE8N3HG8fGGVgTkjNehWgxZxZh+z/IMxcfJx04rh7PwNr+nXNlNY6vpiPZGTY5087pA+c7yH5xmu6gWVYIxcMryhRvZRgE9yB2oA5Bb6SL4nXUx0vUfs/wDZ6wCdbU+WzhyxAPfg1Q1Sxv5NWvrGPT7iT+0b22uorkR/IiKQWDHsRjge9eh0UAVr2G5ntWjs7o2sp+7L5Yfb+B61ifD9WXwTZrI/mOHmDPjG4+a3OO1dJXO+BP8AkT7b/rrN/wCjXoA6KijNJvXOMjPpQAtFGaKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArzL4D/APIi3n/YSf8A9FRV6bXmXwH/AORFvP8AsJP/AOioq5qn8eHo/wBD28J/yK8T60//AG49B1X/AJA17/17yf8AoJr5y8If8jlpH/X1H/Ovo7VP+QPef9cH/wDQTXzj4QP/ABWWkn/p7j/nX0GA/hVP66HxeYfxqf8AXU+mK8z+F/8AyPXj7/sJD/0bPXpleZ/C/wD5Hrx9/wBhIf8Ao2evnqv8Wn6v8j7TAf7hi/8ADD/0tHptFFFdR4gUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWBN/yUS1/7Bsn/AKMWt+sCb/kolr/2DZP/AEYtAG/RRRQAUUUUAFFFFABRRRQAUVgeNNS1PR/DVxqOkNbh7VfMdZ0LBlHYYI5rNfxNfp41OnXNxDZ2WIvKMlpIfOZhkgS/dBz2NAHY1k+Kv+RQ1b/rzl/9ANYmn+KNSuNYtZJlt/7NvrqS2hRVIkTaOGJzg59McVqeMr61s/COp/a7mGDzLSVU81wu47TwM9TQBpaV/wAgey/690/9BFUfGH/Ik61/14zf+gGr2lf8gey/690/9BFUfGH/ACJOtf8AXjN/6Aa3w/8AGh6r8xPYq/D7/knOg/8AXjF/6DXR1znw8/5JxoH/AF4xf+g10dGI/jT9X+ZUt2FFFFYCCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDA07/keta/697b+T1v1gad/yPWtf9e9t/J636ACiiigAooooAK53wJ/yJ9t/wBdZv8A0a9dFXO+BP8AkT7b/rrN/wCjXoA2NSleDS7uWM4eOF2U+hCk14D4S8QaofHFjK15KxurlUlDMSGDHnivfNVP/EnvDnH7h+T/ALpr538Ht/xXGmHevN0vOOvNergYp0ql10PIx8mqtOzPpMUtIOtLXlHrhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV5l/wofwx/z/at/wB/ov8A43XptFZVKUKnxq53YTMMVgr/AFebjfe3l/w55Xe/A7w1b2FxOl9qpaOJnAM0eMgZ/wCedeT6DoNtqmv2NjcSSrFcTLG5RgGAJ7ZBr6g1X/kDXv8A17yf+gmvnHwh/wAjjpH/AF9R/wA67sHl+FnTm5QTsYY3iXN6dWCjiJK56b/wojwx/wA/2rf9/ov/AI3XT+DvAemeCftn9lT3c32zZ5n2l1bGzdjG1R/eNdLS158MPShLmjHU7cRnGPxNJ0q1VuL3X4hRRRW55YUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWBN/wAlEtf+wbJ/6MWt+sCb/kolr/2DZP8A0YtAG/RRRQAUUUUAFFFFABRRRQBj+KdCm8R6FNpkWoyWCTjbJJHErll9Pm6VTn8LXl/9ni1PXZrm0hdHaAW8ab2TkHcBkcjPFdJRQBzln4PjtNcW9+3TSW0UzzwWZRQsTuME56n2B6Va8WwxTeEdV82JJNtpKV3KDg7DWzWT4q/5FDVv+vOX/wBANAFvSv8AkD2X/Xun/oIqj4w/5EnWv+vGb/0A1e0r/kD2X/Xun/oIqj4v/wCRK1n/AK8Zv/QDW2H/AI0PVfmBU+Hf/JN9A/68Yv8A0Gukrmvh1/yTbQP+vGL/ANBrpaeJ/jz9X+Y5bsKKKKwEFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGBp3/I9a1/17238nrfrA07/AJHrWv8Ar3tv5PW/QAUUUHoaACiuUvfF9xa6zNElkj6dazx29zcmXDK7jOVXHIHfml1nxZPpviy10WGOxHnw+YZbu6MWfmxtUbTk8+tAHVVzvgX/AJE+2/66zf8Ao162bu+t9PtWuL+eO3hX70kj4UfiaxPAEiTeCrOSFw8bvMyspyGBlfkUAbGqBDpN2JSRGYX3EdhtOa+efBy2x8daaJHcRC6Xy2A5JB+XNfQms/8AIDvv+veT/wBBNfOfhH/kctH/AOvuP/0KvWwH8Kp6f5njZh/Fp+v+R9M96WkFLXknshRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAVNTG7SLwZxmBxk/7pr5z8IJnxppS7hxdJz64NfReq/8AIGvf+veT/wBBNfOPhH/kcNI/6+o/516+A/hVP66HjZh/Gp/11PpmlpKWvHPZCiiimAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWBN/wAlEtf+wbJ/6MWt+sCb/kolr/2DZP8A0YtAG/RRRQAUUUUAFFFFABRRRQAUVU1HVbDSYVm1S8gtI3bYrzSBQT6c0kur6dDqEVhNfQJdzDdHA0gDuPUDvQBcrJ8Vf8ihq3/XnL/6AatQ6vp1xqMthBewSXkIzJAsgLoPcdRVXxV/yKGrf9ecv/oBoAt6V/yB7L/r3T/0EVS8X/8AIl6z/wBeM3/oBq7pX/IHsv8Ar3T/ANBFUvF3/Il6z/14zf8AoBrah/Fj6r8wKXw5/wCSa+H/APrxi/8AQa6WuZ+HH/JNPD//AF4x/wAq6aqxP8efq/zHLdhRRRXOIKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMDTv8Aketa/wCve2/k9b9YGnf8j1rX/Xvbfyet+gAooooA5G/8I3t1q1w8N5Cmn3c8dxcQtGS5dBjCkHGDgZ71Y8Q6Dquu2TWRnsI7eeLZO5hYyA9yhzge2eldNRQBGkKiFY2UMqgD5hnpWD4EAHg61AGAJZuP+2r10Vch4U1nTNJ8H2X9q6jaWPmyz+X9pnWPfiVs43EZxkfnSbS1ZUYSm+WKuzotY50S+/695P8A0E185+EgT4z0j1F3Hn86901Xxd4bk0e8SPxDpTu0DhVW9jJJ2ngfNXgnhe8tbfxXpU09zFFFHdRs7vIAFGepJ6CvTwFWmqc7yW3c87MMFinVp2py37PyPp8UtYo8Y+GP+hj0n/wOi/8AiqX/AITLwx/0Mek/+B0X/wAVXle0h3PY+qYj/n3L7mbNFFFWcwUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUGgCnqv/ACBr3/r3k/8AQTXzj4R/5HDSf+vqP+dfR2qDOj3g6/uH/wDQTXzl4QB/4TLSQB0uk/nXrYD+FU9Dxsw/i0z6YpaSlryD2QooopgFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVgTf8AJRLX/sGyf+jFrfrAm/5KJa/9g2T/ANGLQBv0UUUAFFFFABRRRQAUUUUAcl8TJrSHwHqAuhueSIrCPLLkv7YB5rmLyWO8vNQigjklvdQvbSfTpRE3+pURhsNjjBWTj3r1SigDzbQ4z/wkGlWawuNRs7y5kvXMZHysOCW7g8Y+ldT42trm58I6j9mvntNlrKz7I1beNp456V0FZPir/kUNW/685f8A0A0AW9K/5A9l/wBe6f8AoIqn4t/5E3WP+vKb/wBANXNK/wCQPZf9e6f+giqnivnwfq//AF5S/wDoBrWh/Fj6oDO+Gxz8M/D/AP14x/yrp65b4ZNu+GOge1kg/Supq8T/AB5+r/MqXxMKKKK5yQooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAwNO/5HrWv+ve2/k9b9YGnf8j1rX/Xvbfyet+gAooooAKKKKACvCvENpBqEfw/s7tPMt7jUrmKVMkbla5jBGRyOD2r3WvEdV/4+vhv/wBhef8A9Ko65cVrSfy/NHuZA3HMItb2n/6RI7HUvhT4Mt9Ku5odG2yRwuyn7VMcEKSP468W8O6RY3/iXTbW7g8yCe5RJF3sNyk8jIOa+mdY/wCQJff9e8n/AKCa+cvCP/I5aR/19x/zr2MDhMPKnNypp/JHz+PzrNIVIKOJqL/t+Xl5nsw+Engj/oCf+Tc3/wAXXDfFrwN4d8M+E7a80PTvstxJepEz+fI+VKOSMMxHVRXtgrzP48f8iLZ/9hJP/RUteHiaNKNGTUV9x9dkeZY6rmNGFStJpvVOTa/M9NoooruPlQooooAKKKKACiiigAooooAKKKKACiiigAoPSig0AVdRdo9LunT7ywuRn1CmvnfwfdSr4602YEb5bpQxx/ePNfQ+p8aVd9P9Q/X/AHTXzr4QI/4TfS/lXH2pfp17V62A/hVPQ8bMP4tM+lKWkpa8g9kKKKKYBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFYE3/ACUS1/7Bsn/oxa36wJv+SiWv/YNk/wDRi0Ab9FFFABRRRQAUUUUAFFFFABRRRQAVk+Kv+RQ1b/rzl/8AQDWtWT4q/wCRQ1b/AK85f/QDQBb0r/kD2X/Xun/oIql4vYp4L1ll6ixmI/74NXdK/wCQPZf9e6f+gipbu2ivLKa2uUEkM0bI6HowIwRWlKShUjJ9Gg3PP/gTqM2o/CewNy+9oJJIV46KrYA/KvRq8R+E3i/T/DuqN4KkTZLJf3GxtwCxgE7QfrXrtpr+m32sXmmWl0sl3ZKrTov8Abpz0r1M1oT+tTqxg1GXvL/C3/noKzi3CT1Who0U3zY/m+dfl5bnpVO71rT7G7sra6ukSW+cpbrn/WEDPFeTGEpO0VcZeoooqQCikVldQyMGB6EHNLQAUUUUAFFIzqgBdgoJwMnHNLQAUUUUAFFBIHU4ooAKKKZLNHBE0s7rHGvLMxwB+NG4D6KRWDKGUggjII70tABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYGnf8j1rX/Xvbfyet+sDTv+R61r/r3tv5PW/QAUUUHpQAmaM+9ea6xL/wAVBqV407rqVrqFvFaJ5hGIyOVC5wQeSak8aSrLqurvPcPHJY6Ws+mhZCuZSWyVAPzEEKMc9aAPRxXiGq/8fXw3/wCwvP8A+lUdewyPfnSo2sxC12UU4nJC5xznHNecxeBbnxn4R8Pzrqx0q606W5cPDCX+dps5U7lIwU4NYYiMpU2oq70/M9bJ69KhjY1K0uWNpK+r3i0ttd2ej6upbRr1QCSbeQAD/dNfOvhCN28a6UoRiVu0yAORg85r0o/C7xMRg/EjViP92X/49VKD4LanbXRubfxtdw3DdZY7dlY/iJc104fHVaMJR9k9fNHPicmwFecZfXIq39yf+R6yOteZ/Hj/AJEWz/7CSf8AoqWj/hV/if8A6KTq35S//HqrX/wc1jVYFg1Tx5fXsStvEdxC8ihsEZwZTzgnn3rzqzrVKbgob+aPcy2nl2DxdPESxSai725Z/wCR6tRRRXefKhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRQaAK2oI0mm3KIu5mhcBfU4PFfPPg/T7o+OtPi+yv5kNwGdcH5ADzmvo/FQpaQRzGWOGNZG6uEAJ/GuqhiXRjKNtzjxGGVacZX2JaWkpa5TsCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACsCb/AJKJa/8AYNk/9GLW/WBN/wAlEtf+wbJ/6MWgDfooooAKKKKACiiigAooooAK5PxreaxZ3OnyafLfwaeqTNey2MEcjpgLtJDg8fe6V1lZms6HHrcaxzXt7bR4KultLsEgPUNwcigDkrTxLqN3dR6tb6iZdOOqppi2vlLh1OB5hOM7snOM49q6Dxvqdpp3hHUftjsnnWsqJtjZ8naf7oOPxpLfwVpNpqUV3bidEiYSLaiT9zvAwH2/3vfNW/FX/Ioat/15y/8AoBoAt6V/yB7L/r3T/wBBFW6qaV/yB7L/AK90/wDQRVugD5n8ZeE9IHi7VftN/c6Zq8+oDyJSh8tYifvgjB6e9aug/DY6fcX1tpXxGjj1G5g8uYIQRIT05PXj05Fe66lommaxC0ep2MFyrDB8xMnH161kf8K78KCNkj0W3j3fxJkMPoc8V6NHOs1oUlQhNOC2TSei2WzPWqSyvEfvKkZxm97Wab76289PxPH4/hvr93rklrrfxChtp5I1E8UExDMgG1e4BOBXQw/DPUC2mifx8k11phzpqbUwjZ+XI6t713CfDLwqLkTyab5r4xmSRm/HrTW+GHhdt5Fk6O3KuszBk/3T2rWfEGaT/kXol6fy9tDH6tlfSc//AAFb/wDgRyN5ZfFSBrkQeLNG3yOAquqLsJH3RkcY/XNaUHh/4hLpLWb+N7BpQuJna1BaLPv/AI1tw/C/w1FG6vBcTM4/1ktwzMD/AHgfX3qKX4V+H5SD5moK2MOy3bAyf73rWEs3xTSSpU1/26v8g+rZdf8AjT/8B/8AtjB0HwR4+0PSbTS7LxZZQ2lmWKSG28xpgTnDbugHtWjHpnxFkhltx4q0l5G/5ai1y8QPoBx+dXn+F2iPcF/tOpqhGPKW8baKP+FW+HxEFSTUYznJdLxgzexPpSnm2LqScpU4NvX4Y/5AsLl6X8aX/gP/ANsYqaX8S7bWrfzPFmkz4QqYJIAvme+0c5+lXJtN+JN8g2eJdJsSGLfubbfntt5zxWg/wv8ADzszn7b5pIIl+1NuXAxgHtnvSj4YaArZEmoAZBx9sbFJ5piW7+yp/wDgK/ysH1fL7fxpf+A//bHIa34I8d69o0Wmaz4w09RDcrOtykeyTf2XggcHpU2neB/iHFEgtviNFIsamJSbcSZGe5PU579a6iT4X+Hpd2/7adw/5+m4Oc5+tMn+FmhOVNtPqFoVHAhumAz649a1/tvHKPIoQtvblj1/7dBYXLn/AMvpX/w//bFS80b4g/2c8Mvi3TIQw2m4+ybWA6Z64zVG60H4lT21tZw+K9PCREt9rWHEkwxjDDp75Fbq/DLRTuF1PqF0rfwTXTED8Kh/4VToGBi41RcDAxetxWUc0xUdqUP/AAGP+QfVsv2daX/gP/2xh3/gzx5eabp1lfeMrSNbOdZVu1gKyOy5AB5w3Xoa07vQ/iC1s8f/AAmenxbl2+Z9hAI9/TNaI+GmhlUWSXUJY05CSXbEZ9frUd58LtBuwMy6hFj+5dtz+dOWbYtpfu4af3Y//IijhsvvZ1pf+A//AGxSGl/EaSERw+J9J3Q4Vitrkvj+9zxnrWTq3hPxxr2kz+HdU8Yab5UsWHK2+JZF3ZyRnjB4/Ct+D4V6LBEyre6oGf77LeMN31pJfhP4eluxcebqCNtC/LdsOPr70oZti4NShTgmnf4Y/wCRf1XLrtOtK3+H/wC2Ktro3xB0y2tba28RaTcwwIEUy2xBfAwFOD+tTTaf8Q/tEdyfEOkQ9R9k+ynY3vuJzkVb/wCFY6F5YQTaiADkf6Y3B9aF+GWhi3eKSbUZQwwrSXbFo/dT2qXmWIbu6UL+i/yI+r4D/n9L/wAB/wDtjCOg/EOz1sa1L4n0x/Lh8k20sZWGQddx54P0q/Z2XxLt5ZZZNT0W5E58zy5IWxD/ALKkHke5q5/wq/Q5LUwXU+oXC8BTJdMSo9BSj4Z6Us6yx6hqyMowcXjc+lVLNMTL4qUO2y/yH9WwHStL/wAB/wDtilqMXxGl0+4D6hoVijoQZ0R8wj+8CTj86jsIviK+mWcA1fQ/LCqBfBGd5cDjjODnrxV+T4YaTNFLDNqGrPBL9+NrwkN9eKji+E+gQ7RHcamFT7ii8bC/QVH9pV7cqowt6L/IpYbL7a15X/w//bELWHxJlZca9o8KRMSzLbE+Z/snJ4/Cpg3xFZjCbjw+kijIwkhYj6bqk/4VdoflhBdaoO5IvW+Y+9PT4ZaJHGdlxqQlJz5/2xt+PTPpQ8wrPejD7l/kT9XwH/P6X/gP/wBsVWsfiPEXlOt6OwZR8jWzAR+4Oe/vVaKz+KCXPnHWNFn3nm1aEhYx2wRya1m+G+jSKRPcalNnqXvGpr/DXRvLT7NcahbyJ/y1ju23H6561SzKuv8AlzD7l/kT9XwP/P6X/gP/ANsYuk6R8S9JjvVXWNJ1R7iUyeZcq+YSR91QDgD2rRiX4lONks+gRkDG4JISffGalPwx0gf6q+1WIFtzBbxvmPrUqfDbRo8GO51NW7sLxsmiWZV5u8qML+i/yH9WwK2rS/8AAf8A7YypY/iLHqaTLrmhPH5RH2Zoyqtz97rn264piD4j2d3cST6roLfanDJDLuAgA/hXnJB96vRfCjQlR1uLjUJ+oRmuTmNc5wDUk/wr8O3Hllzfbo+jC6bJ+tL+0sRa3sYfcv8AIv6tlyf8aX/gP/2wBviPu5/4R4Z6Ltl/P71QyL8SYi+LvQWLYIDRuNg9uatP8NNIefzvt+rCQDAYXrVE3ws0WTJlvdWckYyb1qn+0Kq2ow+4lYbA9a8v/Af/ALYYx+JLRhDL4fjY9GCyEt7daPM+IkYHmXHh3dwNrLIP/ZqkX4XaMsu/7bquVP7v/TG+T6Ur/C3QHRA0moMVYkubtizfU0f2hV/58w+4f1fAX/jy/wDAf/tjGs9P+JFibye98SaOqTSltsseVt84wFPH5GrkFn8Sv3if25osxZ927yDmMf3cA9D781dk+Fnh+W1W3ke+ZM5cm6YmT0z64pY/hfoMTK0U2pIyjGReNyfU+pq5ZpiJO7ow+5f5C+rZfbStK/8Ah/8AtiJIviQGJF5oMoHrE/Ptwe1CL8SFjKtcaC+SSZSkg2+2M1Kvwx0dQwN7qp3/AHv9MYZOetJB8M9NhV4jqWqvbt/yxN2cfn1qP7Qrf8+YfcH1bA9K8v8AwH/7Y2vDaa8lrN/wktxZzys+YmtFKqF9Oa2ayfD/AIdtPDdpJbWEtzJG77v9IlMhB9ia1qwlUdR88kk30Wxx1IwjNqm7ru1b8NTA07/keta/697b+T1v1gad/wAj1rX/AF7238nrfqTMKKKKAKU2k2E9+t7NZQPdIMLM0YLAfWmz6LptyYDPYW8pt/8AUl4wfL+npV+igBuK5+PwRo8KlYftsSlmbYl7KoBJJOAG45Jroqr387W+n3E0Yy0cbMB7gUAY/wDwh2l/89dQ/wDA+b/4qqNt4MhGtXxuJr/7GVj+zj7fLwcHd/F64qr4UvL5dTsFudQnvV1LTRdyLK+RG+R930U5xj2rKTXtVtY7TWPts07ag94jWjMPLiEaOU2jsQUGfrQB13/CG6X/AM9dQ/8AA+b/AOKo/wCEN0z/AJ66h/4Hzf8AxVZnha4vINcSzuNQnv47nTorwmdt2x2ZgdvopxwPauyoA5//AIQ3Sx/y11D/AMD5v/iqP+EO0v8A566h/wCB83/xVbdzIYreSRV3lEJC+vHSuG8K6hqH9q6U1xqE14uq2bzyxysCsbhuCg7DtigDQ07wZCLm/wDt01+YzcE23+ny8R7V/wBr1Bq//wAIbpf/AD11D/wPm/8AiqzibqPx1FbaZqF5cqFd7+OV90MKkfIB/dbPT2BqDwLNqNnIumeJZNQOrtbCV/tNwJY5BuwWTH3fpQBsf8Ibpn/PXUP/AAPm/wDiqP8AhDdL/wCeuof+B83/AMVW/UczFImdV3FVJA9aAMP/AIQ7S/8AnrqH/gfN/wDFVn6f4LjXUdTN7NftbtOptB9vl4Ty1z/F/e3VS8NahftqulzXGoTXS6rHO0sLtlIijHBQdgMAU5brUtH8VXM2stqBt7q4ZLDFyGgPyZCmMDIOQeaANv8A4Q7S+nm6h/4Hzf8AxVL/AMIbpn/PXUP/AAPm/wDiqwvDF/f/ANr6Y9zqE14uqWsk0sUjZWJlfgqOwxxj1rvKAMD/AIQ3S/8AnrqH/gfN/wDFUn/CHaX/AM9dQ/8AA+b/AOKrV1W4e00m6uI13PFEzKD6gVyXhG8vl1K1gudQnv0v9MjvXMrbvLkOM7fRTu4HtQBZtPBkI1bUWupr9rQmP7KPt8vA2/N/F/erQ/4Q3S/+euof+B83/wAVWN9lvm8b3lvpuq6hLHa2LtJHNPuj86QfuxjHGAc1Z8E3F3Ar6Zrbah/asdvFJOt3cCZWyMFkIGACQeKAND/hDdM/566h/wCB83/xVH/CG6X/AM9dQ/8AA+b/AOKrfpG5WgDA/wCEO0v/AJ66h/4Hzf8AxVULbwZCNcv2uJr82Rjh+zj7fLww3b/4v92p/CRuk1TX7O6vbi8S2vEETXDbigMasQPQZPSue/tzVLVoNYF5NM11d3EDWTN+6VVRiu0djlc596AOp/4Q7S/+euof+B83/wAVS/8ACG6X/wA9dQ/8D5v/AIqsbwje3yatawXOoTX6X2mpeOZTkI5PJX0U9h7V3FAGB/whul/89dQ/8D5v/iqT/hDtL/566gf+3+b/AOKrdmTzIXQOyFgVDKeVz3HvXNeCXuZNJ1CC8vJ7trfUZ4Elnbc+1TwCfagCvpvguJb7Uzfz35ha5BtAL+XiPy1/2v726tH/AIQ3S/8AnrqH/gfN/wDFViD+0n1XVrTQNTvLxY7GVHkuJQyx3RH7sIcYGM8/hV/wbdTL9p07UDfjUbdY2nS9nE3VeqsBjHtQBc/4Q3S/+euof+B83/xVH/CG6X/z11D/AMD5v/iq3x0psi7o2XcVyCMg8igDB/4Q7Sv+et+f+3+b/wCKqjZeDIRqWoG7nvzbF0+yj7fLwuwbv4vXNWPBrXKwatbXd5cXn2XUZIY5bhtzbQqkDP4mufh1LU9Bnv49Ze/+3XEdzNZebcCSFkU8YUD5cAr1oA6f/hDdL/566h/4Hzf/ABVH/CG6X/z11D/wPm/+KrN8KXF3Fq5sri/nv4prCK73zvuKO2MgegOcge1diKAMD/hDdL/566h/4Hzf/FUn/CG6X/z11D/wPm/+KrfYblIzjI6jtXLeDmu1vNfs7u9ubxbTUfLie4bcyqYo2xn0yx4oAhsvBkQ1rUWuZr/7E3lfZR9vl4+U7/4vXFaP/CG6X/z11D/wPm/+KrJMF3/wn8ltp+q37x29m808Us+6PzHyEXGOPWn+Cbi8tpG0vXm1D+1lto5ZRd3ImRhyCyEdOQePpQBp/wDCG6X/AM9dQ/8AA+b/AOKo/wCEN0v/AJ66h/4Hzf8AxVb4prjcpGSuRjI7UAYP/CHaX/z11D/wPm/+KqhpnguJbjUPt81+UNyTbYv5eI9q4/i9c07w8bj/AISm9htNQur/AE2CILLJdPvAn3cqh9h1rBOu6pbywauL2aQ3V9c2zWTH92iJG5G0f3soDn3oA6z/AIQ3S/8AnrqH/gfN/wDFVPp/hnTtM1A3tqLhrjyzFvmuHkwpOcDcTjkVg+Fru+TWLSG41Ce9S/04XbiVt3lvuP3fRe2K7UUAA4FLRRQAUUUUAFFFFABRRRQAUUUUAFZ3iG3lu/DOpW9shkmltZERB1ZipAFaNFAHM2PiC8t9PtoZPDerbo4lVsInUAD+9TrnxdLaWstzceHtWSKFC7t5acADJP3q6SsrxJe21joNw17C08coEPkocNIXO0KPzoApx+KZ5Ylkj8OasVdQyny05B/4FTv+Emuv+hb1b/v2n/xVWvD+rR6nYugtntJbRzBLbyEExleOo4I961aAMD/hJrr/AKFvVv8Av2n/AMVR/wAJNdf9C3q3/ftP/iq36KAMD/hJrr/oW9W/79p/8VR/wk11/wBC3q3/AH7T/wCKrfooAwP+Emuv+hb1b/v2n/xVI/ii4jRnbw5qwVRkny06f99VvSsUidlQuyqSFHVvauVXxqtzpWX0m4F7LdyWK2PmLud1GW+bpjbk0AWLTxfJe2cN1beHtWeGdBJG3lp8ykZB+96VL/wk11/0Lerf9+0/+KrMt/GVlZ6PaRaTps88cFuxlhVgDaxxHY2c9SCCMDriuttriO7tYriE5jmQOp9QRkUAYv8Awk11/wBC3q3/AH7T/wCKo/4Sa6/6FvVv+/af/FVv0UAYH/CTXX/Qt6t/37T/AOKo/wCEmuv+hb1b/v2n/wAVW/RQBgf8JNdf9C3q3/ftP/iqP+Emuv8AoW9W/wC/af8AxVb9cvqXjVdEvTFrGmz2sLhzBPvVhNtxxgcjORjNAC2PjNtRtFubPw/q0kLMyhvKQZKsVP8AF6g1Y/4Sa6/6FvVv+/af/FVmWHii10nTGsk0ie2vI5lSPTy6lnaUlwQ3Tn5j7YNdDoesRa5pxuYo2hZJXhlifrHIjFWX3wQeaAKP/CTXX/Qt6t/37T/4qj/hJrr/AKFvVv8Av2n/AMVW/RQBgf8ACTXX/Qt6t/37T/4qj/hJrr/oW9W/79p/8VW/RQBgf8JNdf8AQt6t/wB+0/8AiqP+Emuv+hb1b/v2n/xVb9c5rXi9PD9+I9TsXjtGJC3Syq2SFLfc+9j5TzQAy18ZNeiU23h/VnEMrQv+7Th1OCPvVP8A8JNdf9C3q3/ftP8A4qsyx8WWlhaXhfSri0uJHFyluzqzXBlPBBHAye3at/Q9aGswT77Z7W4tpTDPBIQTG2AcZHB4IP40AU/+Emuv+hb1b/v2n/xVH/CTXX/Qt6t/37T/AOKrfooAwP8AhJrr/oW9W/79p/8AFUf8JNdf9C3q3/ftP/iq36KAMD/hJrr/AKFvVv8Av2n/AMVR/wAJNdf9C3q3/ftP/iq365nWfHOn6JrY0+4ilcLC8s0yY2xYGQp9yAfyoAIPGT3FxcQw+H9WaS2YJKPLT5SQCP4vQip/+Emuv+hb1b/v2n/xVZtv4qtrP7dNLpFxbalPLETaM6lpyy4Qg9Bwp+mK39D1hNaspJRC9vLDK0E8LkExuvUZHBoApf8ACTXX/Qt6t/37T/4qj/hJrr/oW9W/79p/8VW/RQBgf8JNdf8AQt6t/wB+0/8AiqP+Emuv+hb1b/v2n/xVb9FAGB/wk11/0Lerf9+0/wDiqP8AhJrr/oW9W/79p/8AFVv1g6l4mez8SQ6Na6c13NJB57N56RhV3bf4up+lAFePxk8t7NaJ4f1YzQKrSL5afKGzj+L2NT/8JNdf9C3q3/ftP/iqz7fxZpseuT3T2UsEF0WgS/ZgUmaHcSoHUY+f8jWloXicazdNby2M1lKYVuIllYHzIicBuOn0oAh0A3d14l1bULjTrmyhmigSMXAUFiobPQn1FdHRRQAUUUUAFFFFABSModSrDIIwQe9LRQBjWHhPSdMleSygeJmI581jtAOQo54X2HFOg8LaRbanNfxWuJ5g27LEqNww2F6DPfFa9FAGXpHhzTdCeR9OhZGkAVizliFHRRk8Ac8e9alFFACEZrGsvCWkafcvcWdu0Uj8ZEjfKM52rzwM9hW1RQBiWvhHS7LU57+1WdJ7h2kl/wBIfa7HqSucVPpnh3T9JupLm0SQzyLtaSWVpG25zgFicDPatSigAoNFFAGLa+EtIsrqS5tYHjlk6kSthcnJCjPygnripF8NacurDUSkklwrFk8yVmVD6hScA+9a1FAGTpnhnS9IvZLqxgMcsgKklyQoJyQoP3Rn0rWoooAa6LIhRwCrDBB7isjTvCmk6S7NYwvGWIP+tY4AzhRk8Lz06Vs0UAUo9JtIpLySOPbJenM7hiCxxgc9uKh0vw/YaPJLJZI/mzACSSWRpGYDoMsSce1adFABTScj+tOIyOa4rxtZ3Ud/Yz6ZdG1k1BxYXRyeY2/iHowzwaANiw8JaVpl7cXdotws1y4klLXMjbmHfBNSw+GdJh1aTUo7bFxIWJy5KgsMEhegJHGa53x34e0s+HbQyWqs9vLDBG5JBCbunWql5q2s6b4ml0bw3Z28FjYCEiJ5EUSBySfvc9u2aAOv0jw3pmhySvplv5TSAKSXLYUdFGegGegrVzXmGu6/rJh1ewudSNrdFXe0WCFHQRrIvzbwcg4PINehaQZP7JtvOuxeP5YzOFCh/fA4oAuSxiWJ42yAwIODg/nWLpfhDStGEv8AZ63EfnSGVw1y7ZY9Tya3KKAMHTPB2k6Qs4sFuIlnDh1+0uRljkkAng57irmlaBYaN5psY3DzEGSSSRndsDAyzEmtKigAFI6h0ZTnDDBwcUtFAGJpnhLS9HlmksRcKZ5PNk3XLtub15NPh8K6VBeyXQheWaRGQmaVpMK33gAx4B9q2KKAMzR/D2naF5n9nQshkwGZ3LnA6AEngD0rToooAQjcpB6EYrE07whpWlXVxcWS3CSXMglmLXLtvYcZIJ9h+VblFAFOLSbSC6vLiKLZNe489wTl8DA+mBVfS/Dun6PNLNZxuZpQFeWWVpGIGcDLE8cnitSigAprxiRGVujDBxTqKAMLSvB2k6KhTTkuIoy27Z9pcjOc5wT61NH4X0qLWH1NLb/SXJYksSoYjBYLnAJHGa16KAMrSfDWmaJNJLp0BjeQbSWcttXOdoyeB7CtWiigAooooAKKKKACiiigAooooAKKKKACiiigArH8T6VPq2j+VZFBcwypNCJDhSysDgntkZGa2KKAMPw1pl7ZJfXWqiJLu+uDM8cLlljHQKGIGeO+BW5RRQAUUUUAFFFFADZC4jYxqGcA7VJwCfrXn1r4d8TpP9vkstPju7bU5b63jF4zLIJUKFWOwYwDnODXodFAHn0fg3WtOtWOn/ZZri/gnivRJIVWIzSM5ZDg7sF2GOM8dK7jTrT7BpltaA7vIiWPd64GM1ZooAKKKKACiiigArgNS8M6/r2p6m2r2tn5UitFYzLdEm3QMGHybOrFRk5rv6KAOEl8Ma7eXT67cx2cerJPDJFbLMTEwjVlIL7cjIdu3HFdD4W0m40jS5Vvihurm5luZRGcqhdy20HvjOM962qKACisHVfGej6PqJs72ZxIu3zGVcrHu6bj2zVa0+IOhXdmboSTww+SJ43mhKeamcZT154oA6eiueHjbSjb7wt15/m+T9k8g+eWxnGzr05qk/j62tvEM9je2txFBHBDKJPKJKeYzL+8H8IyBQB11cR4r8K6h4ov/L+xWNoiblXU1lLTlCpGzbtGASf73au3BBAI5B6UUAcE3hTXL+Nru+W0hvrWKCO0jjlLI/lsGyxwMZI6YOK6Hw1pl7ZDULvVViju9QufPkihcusfyhQAxAzwo7DrW5RQAUUUUAFFFFABXnfiH4Z3Wp33mWetXSwzPNJMkpUgM64GPlyR269K9EooA4SXwzrt/K+rXkdnDqcLwNbwJMWjYR7vvNtBGd7djiuh8MaXdabZ3Umo+Wt3e3T3UqRNuWMtj5QcDOMdcVtUUAFFFFABRRRQAVyni/QbvWpkSz0qwlcoFXUJpSstt82flAU5/MV1dFAHn0HgbU5IrfSL2WA6ZYyzzw3KuTLK0iuMMuMDHmN3Patjw1ourW+qC91tbaNoLNbKJYJC+8A53nIGM+nP1rqaKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAqOW3hn2edEkmxgyblB2n1HvUlFAEctvDcR7J4kkXIO11yM+tV7rSNNvriO4vdPtbiaL/AFcksKuyfQkZFXKKAKKaJpUcs8sem2iyXAImdYFBkB67jjn8as21pb2VukFnBHbwoMLHEgVV+gFS0UAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUm33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAC0Um33NG33NAHIav4HS71q71O2Fi7XSDzUurJZmyBj5GJ+XNZelfD69vPDFhba9fbJbWyWGCNIgDA24N8xz83KgduBXoe33NG33NAHIf8ACF3n27+1jqyHWfOEguPs/wC7wEKbdmfQnvVi78IS3zalLdX6mbULaGF2SLAUxsTkDPfPSun2+5o2+5oAbEnlQomc7VAz64p9Jt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFJt9zRt9zQAtFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAf/2Q==)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UEqYj2vMhvW3"
      },
      "source": [
        "Unlike classification, regression is not a model that predicts one label. Regression is to model and predict the relationship between the dependent variable according to independent variables. From a machine learning perspective, the goal is to find the optimal weight parameter for each data and create a regression equation with the smallest error. Since we have to compare the three models, the error is **MSE loss**.        \n",
        "The models in turn used polynomial, neural network, and K-NN models. It is the same as the learning process in classification, and only the result value is different by mse loss, so the analysis of the code will be omitted.  \n",
        "        \n",
        "Looking at the table above, the loss of linear regression is the smallest. Next, the loss differed in the order of knn regression model and neural network regression model. Looking at the complexity of the model, the more complex the model, the lower the performance. The data is not very big, so a simple model yielded better results.\n",
        "Below, the first figure shows the distribution between the predicted labels and the actual labels.          \n",
        "Below, the first figure shows the distribution between the predicted label and the actual label. It is normal when most of the data are distributed in the red box, but the predicted values are too sparse. The distribution of the MCI group (group 1) overlapped with the normal group (group 0) too much, so we learned and predicted incorrectly.         \n",
        "The next graph figure, it shows the MSE loss per epoch in neural network regression. Looking at the graph, you can see that training did not overfit. Examining the two figures, the distribution of the data is a major factor to poor learning."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XY6720MyMhyA"
      },
      "source": [
        "# Discussion\n",
        "\n",
        "In this project, I used about 1600 subjects' cortical volume and average thickness of 70 brain regions as data to be learned. This dataset was an imbalanced dataset with more than 50% of the MCI groups. In addition, because the regional volume and thickness of the brain are different for each subject, the distribution in each group is also uneven and widely scattered. Since this data distribution is a factor that prevents proper learning, I focused on analyzing and preprocessing the data. The data preprocessing methods I used are ***Resampling (oversampling & undersampling), Removing outliers, and PCA and Pearson correlation methods***. In order to make predictions better, the data were processed in various ways, but as a result, the performance was not more noticeable than the baseline.\n",
        "\n",
        "Then, we solved two tasks: classification and regression. We implemented three models per task and compared the performance of each. **MLP, LightGBM, and SVM models** were used for classification, and **linear regression, neural network, and K-NN** models were used for regression.Almost all 6 models did not train well, perhaps because I didn't extract meaningful data well in the preprocessing stage.         \n",
        "Since it is medical data, I thought that recall is more important than accuracy, so I tried to find the recall rate in the Lightgbm model, which has relatively better performance, but this was also not good at about 53% on average.\n",
        "\n",
        "I think we should consider a model that extracts the correlations and features well in data and classes.\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3RYybDvKPz_z"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}