🎬 🍿unsupervised_learning
- dataset: movielens dataset
A Website to recommend movies based on user input. Non-negative Matrix Factorization returns movie recommendations based on three given movie titles.
- Clone the repository
- install dependencies:
cd app && pip3 install -r requirements.txt
- run the app:
env FLASK_APP=application.py flask run
- open a browser on the specified location
- Exploratory data analysis and data wrangling (for the small movielens dataset) in jupyter notebook
- Python script to create cleaned up dataset
- SQL script to load the cleaned up MovieLens data into a Postgres database
- NMF model training with sklearn
- Recommender function based on user input (3 movies) and NMF model
- Flask web interface for the recommender