Preference-aware community detection for item recommendation

In recent years, researches on recommendation systems based on social information have attracted a lot of attentions. Although a number of social-based recommendation techniques have been proposed in the literature, most of their concepts are only based on the individual or friends' rating behaviors. It leads to the problem that the recommended item list is usually constrained within the users' or friends' living area. Furthermore, since context-aware and environmental information changes quickly, especially in social networks, how to select appropriate relevant users from such kind of heterogeneous social structure to facilitate the social-based recommendation is also a critical and challenging issue. In this paper, we propose a novel approach named Preference-aware Community-based Recommendation System (PCRS) that integrates Preference-aware Community Detection (PCD) for recommending items to users based on the user preferences and social network structure simultaneously. The core idea of PCRS is to build a community-based collaborating filtering model in the user-to-item matrix, so as to support the estimation of users' rating for each item. Based on the social network data, we detect communities through users' Social Factor and Individual Preference for our community-based collaborating filtering model. To our best knowledge, this is the first work on community-based collaborating filtering model that considers both social factor and individual preference in social network data, simultaneously. Through comprehensive experimental evaluations on a real dataset from Go Walla, the proposed PCRS is shown to deliver excellent performance.