Recommendation based on trust and distrust social relationships

Recommender system has become increasingly popular in recent years, since it is an effective way to solve the problem of information overload problem. But it is still subject to some inherent problems, such as data sparseness and cold start. Many studies show that the integration of social network information is a very effective way to solve such issues. The studies on recommendation methods that incorporate social relationships, not only take into account the preferences of the user for the item, but also the interaction between the users according to their behavior and the social relationships. And now, the application of social relationships has extended from the trust relationships to the distrust relationships. While the collaborative filtering is the most important and widely used recommendation method, there is little work on combining with trust and distrust social network relationships. So, this paper proposes the methods of integration the trust and distrust social relationships, TDUCF1 and TDUCF2, and with the improved cosine similarity, to improve the collaborative filtering recommendation algorithm, which combined the users' trust and distrust social relationships, and effectively alleviated the sparseness. The experimental results show that the proposed methods outperform the state-of-art algorithms.