Incorporating user rating credibility in recommender systems

There have been many research efforts aimed at improving recommendation accuracy with Collaborative Filtering (CF). Yet there is still a lack of investigation into the integration of CF algorithms with the analysis of users’ rating behaviors. In this work, we develop an integrated CF-based recommendation solution by incorporating the credibility of users’ ratings, the demographic information of people, and the ontological semantics of items. The users’ credibility values are calculated based on their ratings and they are used in finding credible neighbors to improve the accuracy of recommendations. The demographic information and ontological semantics are used in the similarity measurement of users/items to alleviate the issues of sparsity and cold start in CF algorithms. Experiments are conducted on real-world datasets of MovieLens and Yahoo!Movie. Compared with baseline methods, a set of experiments shows that the proposed approach improves the recommendation quality significantly.