CHANGELOG

See: https://github.com/ocelma/python-recsys/commits/master

2011-05-09
    * Added MAP (Mean Average Precision) metric and (nose)tests

2011-02-01
    * Added external SVDLIBC executable in recsys.utils
    * Load larger matrices using SVDLIBC instead of divisi2
    * SVDLIBC examples

2011-01-23
    * Improved(?) K-Means clustering and added a test

2011-01-12
    * added recsys.utils: compute SVDLIBC [http://tedlab.mit.edu/~dr/SVDLIBC/], and import U,S,V to recsys.algorithm.factorize.SVD
    * added new test_utils.py to check SVDLIBC to recsys import

2011-11-14
    * add_tuple raises more exceptions, if tuple format not correct

2011-11-08
    * better explanation of the format parameter in load_data
    * save shifted matrix only if mean_center is True

2011-10-14
    * loads and saves zip bigger than 4Gb (allow64=True)
    * loading data from file (load_data method) more robust now

2011-10-08
    * Added the whole project at github