Comp 4990 - Project (Team-17)

Web Application for Sequential Recommendation System

Project Description

• This work was dedicated to understand the functionality and applicability of reinforcement learning and using it for a sequential recommendation system for movies. It is based on content-filtering. This project uses reinforcement learning techniques specific to on/off policy learning, to train a model to better recommend films to users using the Movielens dataset. The project will be distributed such that primary focus will be on a static recommendation system and will progress to being dynamic sequential recommendation system.

Built with

• Python
• JavaScript
• TensorFlow and other ML libraries
• Flask
• ReactJS
• TMDB API

Dataset

• MovieLens 1M

Steps to run the Project

1. Clone the repo through https://github.com/Capstone-RL-2023/Web-App-Recommendation.git
2. Then, cd to the backend and run the server.py (python server.py)
3. Open new prompt/terminal window, and cd to the frontend folder and then type "npm start".

Functionality

The frontend will have the following functions.

• Information about the movie recommended will be displayed
• Giving the next recommendation for the user when pressing the "next recommendation" button
• view stats of the choices made
• changing the user id which will change the user the movies are recommended for.