An Unified One Class Collaborative Filtering Algorithm

The problem of the previous researches on One Class Collaborative Filtering (OCCF) is that they focused on either rating prediction or ranking prediction, no concerted research effort has been devoted to developing recommendation approach that simultaneously optimize both ratings and rank of the recommended items. In order to solve this problem, a new unified OCCF approach (UOCCF) based on Probabilistic Matrix Factorization (PMF) approach and the newest Collaborative Less-is-More Filtering (CLiMF) approach was proposed. Experimental results on practical dataset showed that our proposed UOCCF approach outperformed existing OCCF approaches over different evaluation metrics.