This repo contains my recommender systems for anime.
This project contains following recommender system types:
- Content-based using user profiles based on anime genres,
- Content-based using clustering on user profiles,
- Content-based using item similarity scores,
- KNN-based collaborative filtering,
- NMF-based collaborative filtering,
- neural network-based collaborative filtering.
This project was completed using Python, Jupyter Notebook, scikit-learn, pandas, numpy, matplotlib, seaborn, scikit-surprise, keras.
Advantages and disadvantages of different types of recommender systems were considered. See the presentation for more info.
Initial data was processed using text processing methods. Some useful visualizations were constructed, like this wordcloud of anime genres.