A hybrid time-series link prediction framework for large social network

Abstract

With the fast growing of Web 2.0, social networking sites such as Facebook, Twitter and LinkedIn are becoming increasingly popular. Link prediction is an important task being heavily discussed recently in the area of social networks analysis, which is to identify the future existence of links among entities in the social networks so that user experiences can be improved. In this paper, we propose a hybrid time-series link prediction model framework called DynamicNet for large social networks. Compared to existing works, our framework not only takes timing as consideration by using time-series link prediction model but also combines the strengths of topological pattern and probabilistic relational model (PRM) approaches. We evaluated our framework on three known corpora, and the favorable results indicated that our proposed approach is feasible. © 2012 Springer-Verlag.