An Effective Model-Based Trust Collaborative Filtering for Explainable Recommendations

Nowadays, many companies through the world wide web like YouTube, Netflix, Aliexpress and Amazon, provide personalized services as recommendations. Recommender systems use the related information about products or services to suggest the most relevant of them to particular users. The recommendation is usually made based on the prediction of the users’ constraints and interests. Despite that the most of the existing recommender systems give accurate recommendation results, they do not provide explainable recommendation support. Therefore, in this paper we propose a new effective model-based trust collaborative filtering for explainable recommendations that aims not only to improve the quality of recommendation but also to provide an efficient support for explainable recommendations based on trustworthiness modeling. Our solution can both ensure how the item is recommended and why it is recommended. The proposed method is evaluated on Amazon Instant Video dataset in terms of RMSE.