./User_Based_Recommendations/small-games contains actual work, while ./User_Based_Recommendations contains realization of testing algorithm Actual work consists of bbac.py for co-clustering and recommend.py for generating profiles and actual recommendation. compare.py compares frequent recommendation of modificated work algorithms with testing, get_frequent_recommends*.py writes frequent recommendations of corresponding modification or testing into corresponding file ('' for personalized profiles, 'alt' for bicluster averages, 'old' for non-profile and 'sample' for testing) normalize_profiles.py used to make profiles use number of user or product in place of name.