A simple Java library for using, building and experimenting with different recommender systems
Simply import the project in Eclipse to run it.
I wrote this library for a school project. It contains implementations of the following recommender algorithms:
- User - User collaborative filtering with vector cosine similarity
- Item - Item collaborative filtering with vector cosine similarity
- Slope one item-based recommender algorithm
- Funk Singular Value Decomposition recommender algorithm
New recommender algorithms can be developed by extending the BaseRecommender class.
I tested the algorithms on Movielens 100K dataset, which can be downloaded from here: http://grouplens.org/datasets/movielens/
Please refer to doc/index.html for documentation.
I write a blog on algorithms and data structures at http://kartikkukreja.wordpress.com/