An Improved Collaborative Filtering Recommendation Algorithm for Microblog Based on Community Detection

Abstract

With the rapid development of mobile Internet and the explosive growth of intelligent mobile terminals, microblog has become indispensable in many people’s daily life but also bring overload information. A recommender system is needed but the traditional collaborative filtering recommendation algorithm is suffered from two main problems here: data sparsity and user relationship influence. In this paper, we proposed an improved collaborative filtering algorithm based on community detection for microblog recommendation. We apply the community detection algorithm to analysis the structure of the user relationship network in microblog before we do the recommendation. We make some adjustments to help the collaborative filtering algorithm works better in the community based system. Results of experimental evaluation demonstrate that in a microblog network, our algorithm remarkably outperforms the traditional collaborative filtering scheme by enhancing the recommendation accuracy.