CMPTF: Contextual Modeling Probabilistic Tensor Factorization for recommender systems

• We propose to deem the User-Item-Context-Rating information as a tensor.
• We propose 2 novel recommendation algorithms based on PARAFAC tensor factorization.
• We propose to utilize contextual information in contextual modeling manner.

Contextual information has been proven to be valuable factor for building personalized Recommender Systems. However, most existing solutions based on probabilistic matrix factorization in recommender systems do not provide a straightforward way of integrating information such as ratings, social relationships, item contents and contexts into one model simultaneously. In this paper, we deem the given data as an User-Item-Context-Rating tensor and introduce a high dimensional method of Collaborative Filtering named probabilistic tensor factorization (PTF) which is a generalization of probabilistic matrix factorization. Then, we further extend PTF to a new model named Contextual Modeling Probabilistic Tensor Factorization (CMPTF) which systematically integrates topic modeling, social relationships and contexts in contextual modeling manner to further improve the quality of recommendation. Comprehensive comparative experiments conducted using real-world datasets demonstrate the superiority of our approach.