🛍️ Sentiment-Based Product Recommendation System

Intelligent E-Commerce Personalization Engine

Sentiment-Based Product Recommendation System is an end-to-end ML project that combines collaborative filtering with real-time sentiment analysis. It analyzes 30,000+ reviews to recommend products that are not only statistically relevant but also highly positive in sentiment.

✅ Hybrid Recommendation Engine | ✅ Sentiment Classification | ✅ MIT Licensed | ✅ Flask Interface

🎬 Showcase Gallery

🏠 Recommendation UI	📊 Project Architecture
	(See Architecture Section)

🏗 Architecture

The system uses a two-stage hybrid architecture: first filtering for relevance, then refining by sentiment.

graph TD
    A[User Input] --> B[Recommendation Engine]
    B --> C[Collaborative Filtering]
    C --> D[Top 20 Recommendations]
    D --> E[Sentiment Analysis Model]
    E --> F[Multinomial Naive Bayes]
    F --> G[Top 5 Positive Recommendations]
    G --> H[Flask UI]

Loading

Core Components

• Recommendation Engine (model.py): Implements user-based and item-based collaborative filtering to generate a broad list of candidates.
• Sentiment Engine (mnb.gz): A Multinomial Naive Bayes model trained on 30k reviews to classify sentiment as Positive/Negative.
• Web App (app.py): A Flask-based interface for surgical username input and real-time recommendation display.
• Data Hub (capstone.ipynb): Exhaustive EDA, cleaning, and model-building notebook for the entire pipeline.

🚀 Getting Started

1. Install Dependencies:

   pip install -r requirements.txt

2. Run the App Locally:

   python app.py

3. Explore the Data: Open capstone.ipynb in Jupyter to see the model training and validation metrics.

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Built with ❤️ for Intelligent Personalization.