🍽️ Restaurant Reviews Sentiment Analysis Project

Welcome to the Sentiment Analysis project where we analyze customer reviews for a restaurant and predict whether they are positive or negative. This project leverages Natural Language Processing (NLP) techniques and is deployed using Streamlit for a user-friendly interface.

Project Overview

This project analyzes a dataset of restaurant reviews. The dataset contains:

• 4181 rows
• Two columns:
  • Review: The text of the customer's review.
  • Liked: A binary column where 0 represents a negative review, and 1 represents a positive review.

The goal of this project is to classify customer reviews as positive or negative using a [Logistic Regression, SVM, DecisionTree, Naive Bayes, Random Forest, and KNN ] models.

---

Steps Taken in the Project

1. Data Exploration & Cleaning

• Initial Exploration: We explored the dataset, checking for null values and basic statistics.
• Missing Data: Removed rows with missing values.
• Word Frequency: Identified the most frequent words in the reviews, with “food” being the most common.

2. Data Preprocessing

• Stop Words Removal: Using NLTK, we removed common stop words like conjunctions and prepositions that don’t add much meaning (e.g., "is", "at", "the").
• Special Characters: Applied Regular Expressions (RE) to clean the text by removing punctuation, numbers, and special symbols.
• Processing All Reviews: We applied a loop to process the entire dataset using the steps above for each review.

3. Feature Extraction

• CountVectorizer: We used this technique to convert the text data into numerical values based on word frequency for further analysis.

4. Modeling

• Data Splitting: Split the data into training and testing sets.
• Logistic Regression: Trained a Logistic Regression model to predict the sentiment of the reviews.
• Model Accuracy: Achieved a models best accuracy is 94.3% on the test set.
• Confusion Matrix: Evaluated model performance using a confusion matrix to understand the number of true positives, true negatives, false positives, and false negatives.

5. Deployment

• Streamlit Interface: Created a simple user interface using Streamlit.
  • The user enters a review, and the model predicts whether the review is positive or negative.

---

Model Performance

With the Logistic Regression model, we achieved -->> Accuracy: 94.3% With the SVM model, we achieved -->> Accuracy: 94.2% With the DecisionTree model, we achieved -->> Accuracy: 94.1% With the Naive Bayes model, we achieved -->> Accuracy: 93.9% With the Random Forest model, we achieved -->> Accuracy: 93.8% With the KNN model, we achieved -->> Accuracy: 91.1%

• Confusion Matrix: Provided insights into the classification performance.

---

---

Contributors

• Mahmoud Soliman
• Yara Mohamed
• Mariam Ismail
• Mona Abdelrazek
• Moaz Amen

---