Amazon Hybrid Recommendation System

Developed a Hybrid Product Recommendation System using Amazon Reviews data to generate personalized product recommendations through Machine Learning and Data Mining techniques.

**Key Features:**

- Built a hybrid recommendation approach combining collaborative filtering with product information.

- Implemented K-Means and DBSCAN clustering to identify similarities and patterns in user preferences.

- Applied Autoencoders for dimensionality reduction and improved feature representation.

- Improved intra-cluster similarity compared to clustering on raw features.

- Processed and analyzed Amazon product reviews and user-product interactions.

- Compared clustering approaches and evaluated recommendation quality.

- Used data visualization to analyze clusters and recommendation performance.

**Technologies:** Python, PyTorch, Scikit-learn, Pandas, NumPy, K-Means, DBSCAN, Autoencoders, Collaborative Filtering.

The project demonstrates how Machine Learning, Deep Learning, and clustering techniques can improve product discovery and personalized recommendations in e-commerce applications.