This project implements a movie recommendation system on the MovieLens 100K dataset using four different algorithms:
- Collaborative Filtering with K-Nearest Neighbors (KNN)
- Matrix Factorization-Based Collaborative Filtering
- Similarity-Based Collaborative Filtering
- LlamaRec: Two-Stage Recommendation with Large Language Models (LLMs)
We test our algorithms using five different scenarios that play a cruicial role in determining the performance of a movie recommendation system:
- Recommend movies to a new user (cold start)
- Recommend users for a new movie (cold start)
- Recommend movies to an existing user
- Recommend users for an existing movie
- Predict ratings that a user might give a movie
-
Start the back-end server from the root of this repository using the command:
python CODE/main.py
-
Using a local server (such as HTTP Simple Server), open the CODE/index.html file.
- Numpy
- Pandas
- Flask
- Scikit-learn