Incorporating Context Correlation into Context-aware Matrix Factorization

Context-aware recommender systems (CARS) go beyond traditional recommender systems, that only consider users' profiles, by adapting their recommendations also to users' contextual situations. Several contextual recommendation algorithms have been developed by incorporating context into recommendation algorithms in different ways. The most effective approaches try to model deviations in ratings among contexts, but ignore the correlations that may exist among these contexts. In this paper, we highlight the importance of contextual correlations and propose a correlation-based context-aware matrix factorization algorithm. Through detailed experimental evaluation we demonstrate that adopting contextual correlations leads to improved performance.