Machine Learning Projects Portfolio

Welcome to my CODSOFT Machine Learning Projects Repository! This is an ML portfolio from my Machine Learning internship at CODSOFT. It includes three practical ML projects that demonstrate the use of supervised learning techniques on real-world datasets. Each project involves data preprocessing, model training, and evaluation using Logistic Regression and other essential tools.

---

📌 Projects Overview

1. 🎬 Movie Genre Prediction

• Goal: Predict the genre of a movie based on its description or text-based features.
• Approach:
  • Used TfidfVectorizer to convert movie descriptions into numerical feature vectors.
  • Trained a classification model to predict genres from text data.
• Highlights:
  • Great for NLP beginners.
  • Hands-on use of TF-IDF and Logistic Regression.

---

2. 💳 Credit Card Fraud Detection

• Goal: Detect fraudulent transactions from customer credit card transaction data.
• Approach:
  • Combined TfidfVectorizer for categorical text features (like merchant, job, etc.) and StandardScaler for numeric features.
  • Used scipy.sparse.hstack to merge both types into a unified feature set.
  • Trained a Logistic Regression model.
• Performance: Achieved approximately 99.5% accuracy.
• Highlights:
  • Demonstrates hybrid data handling (text + numeric).
  • Realistic fraud detection scenario.

---

3. 👥 Customer Churn Prediction

• Goal: Predict whether a customer will leave a bank (churn) based on account and demographic data.
• Approach:
  • Text features like geography and gender were vectorized.
  • Numeric features were scaled using StandardScaler.
  • Combined features and trained a Logistic Regression model.
• Performance: Achieved over 83% accuracy.
• Highlights:
  • Focuses on customer behavior modeling.
  • Illustrates simple preprocessing and classification techniques.

---

🧰 Tools & Libraries Used

• pandas for data handling
• scikit-learn for preprocessing, modeling, and evaluation
• scipy for handling sparse matrices
• TfidfVectorizer for text feature extraction
• StandardScaler for feature scaling
• LogisticRegression for classification

---

🚀 How to Run

1. Clone this repository.

2. Navigate to the project directory you want to explore.

3. Install dependencies:

   pip install -r requirements.txt

---

📊 Summary

These projects collectively demonstrate key aspects of machine learning workflows, including:

• Data Preprocessing: Handling both text and numeric data, scaling, and vectorizing.
• Feature Engineering: Converting categorical data into numerical format and combining multiple data types.
• Model Training & Evaluation: Applying Logistic Regression to predict outcomes and measure performance.

Each project showcases practical applications of ML with real-world data, offering valuable experience in fraud detection, customer behavior analysis, and text classification — making them ideal for learners entering the field of data science and Machine Learning.