Modeling Temporal Dynamics of User Interests in Online Social Networks

Abstract Recent years have witnessed an explosive growth of Online Social Networks (OSNs), which serve as a fertile ground for research such as, characterizing individual and group behaviors, identifying information diffusion patterns, and building new recommendation system. This paper explores user interests in social network. While user interests has been extensively studied as the fundamental solution, it neglects the point that a user may change her interests due to social status shift over time. In this paper, we explore two main problems: how user interests change over time and whether user interests have hierarchy. To this end, we first formulate the user interests problem, then adopt semantic enrichment method to determine user interests, and finally employ the topic hierarchy tree model to capture user interests change over time and identify interest hierarchy. Experimental results demonstrate user interests can be divided into primary interest and secondary interest. the primary interest of user hold stability in a long-term period; the secondary interest, however, is more likely to keep up with hot topics or events in the moment. Meanwhile, We also test and compare our model with two existing systems - Who likes what? and TUMS, the result shows that our model can be profiled a more fine-grained user interests.