• Author : Toshihiro Kamishima
• Copyright : Copyright (c) 2012 Toshihiro Kamishima all rights reserved.
• License : MIT License

kamrecsys is a python package of algorithms for recommender systems.

This package targets experimental algorithm used for research purpose. We concentrate on the flexibility of data structure or easiness for implementing new algorithm rather than scalability or efficiency of algorithms.

First, generate sample data sets that you need. Read a readme.md file in a data directory. You then build and install by using a setup.py script.

We tested this module under the following packages:

• Python 2.7.x
  • work on Python 3, but not fully tested
• NumPy
• SciPy
• scikit-learn
• six

• Matrix Factorization

  • Probabilistic Matrix Factorization

• Topic Model

  • Probabilistic Latent Semantic Analysis (Multinomial)

• Programming the Collective Intelligence by Toby Segaran

  • sample_movies

• Movielens Data Sets

  • Movielens 100k Data Set
  • Movielens 1m Data Set

• SUSHI Data Sets

  • Sushi3b Score Data Set

• Flixster Data Sets

  • Flixster Rating Data Set