CiteGeist is a research paper recommendation system that uses content-based filtering via cosine similarity between vectors created with TF-IDF and saved in MongoDB. We deploy this system on a backend which is accessible through the website CodeGeist.
Visit CiteGeist (might face request delay due Render's free tier policy that causes the backend to spin down during low activity).
cd Backend; pip install -r requirements.txtpython manage.py runservercd Frontend; npm installnpm run dev
- Content-Based Filtering: This method recommends articles by analyzing their content rather than tracking user behavior. It finds relevant articles based on what’s written, ensuring quality recommendations.
- Article Ranking: Articles are ranked based on relevance and quality, with the best ones highlighted. Less relevant articles are still considered but given lower priority.
- Text Analysis: The system scans articles for keywords and context to understand their main topics, ensuring accurate recommendations.
- No User Tracking: CiteGeist focuses entirely on the content itself, not on tracking your reading habits or personal data.