About the project
SmartRecommender is a full-stack e-commerce platform that delivers personalized product recommendations by combining collaborative filtering, content analysis, association mining, fuzzy logic, sentiment analysis and probabilistic modelling. Every recommendation method is implemented from the ground up on top of NumPy, pandas and scikit-learn rather than pulled in from a ready-made recommender library, which makes the repository as much a study of the algorithms as it is a shop.
What it demonstrates
- Collaborative Filtering
- Content-Based Filtering
- Association Rules
- Fuzzy Search
- Sentiment Analysis
Stack
- Python
- Django
- React
- PostgreSQL
- Docker
Screenshots
Demo
