Abstract
Conference Title: 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) Conference Start Date: 2018, July 8 Conference End Date: 2018, July 13 Conference Location: Rio de Janeiro, Brazil Context-Aware Matrix Factorization (CAMF) models play an important role in today’s recommender systems due to their effectiveness in solving the rating prediction problem. However, existing CAMF models ignore incorporating the degree of importance of relevant contextual dimensions and interacted subsets of dimensions in the rating prediction process. As a consequence, these models cannot fully capture the influence of relevant contextual dimensions and their interaction on the rating, and furthermore cannot obtain a better recommendation performance. Therefore, our ultimate aim is to propose an improved CAMF model based on fuzzy measures that express both individual and joint contextual dimensions importance. The proposed model is composed of two strategies involving the relevant and dependent contextual dimensions as well as their corresponding fuzzy measures into rating prediction process. The effectiveness of our proposed model is comprehensively tested on three contextual datasets. Experimental results are encouraging and better than those previously reported.