Friend Recommendation in Online Social Networks Combining Interest Similarity and Social Interaction

With the explosive popularity of online social networks (OSNs), the considerably large number of online users and their diverse activities have posed great challenges on social recommendation. However, most conventional methods for recommending friends in OSNs cannot simultaneously satisfy the requirements of accuracy and timeliness. By taking full advantage of latent Dirichlet allocation (LDA), in this study, we designed a friend recommendation approach combining interest-based features and interaction-based topologies with linear time complexity. The experimental results obtained using a real-world micro-blogging network demonstrated that the proposed hybrid scheme outperforms the other three state-of-the-art algorithms in terms of effectiveness and efficiency.