Recommendation with diversity: An adaptive trust-aware model

Recommender systems have become an integral and critical part of various online businesses to achieve better user experience and drive customer and revenue growth. Recommendation accuracy and diversity are important criteria to evaluate recommender system performance. Many different strategies have been developed in existing literature to balance the trade-offs between accuracy and diversity. However, those methods often focus on a one-size-fit-all trade-off strategy without considering each individual user' specific recommendation situation, which leads to improvements only in individual diversity or aggregate diversity. In addition, the trust relationships among users have not been studied to improve the trade-off strategy aforementioned. In this paper, we propose an adaptive trust-aware recommendation model based on a new trust measurement developed using a user-item bipartite network. We show via experiments on three different datasets that our model can not only balance and adapt accuracy with both individual and aggregate diversities, but also achieve significant improvements on accuracy for cold-start users and long-tailed items.