Encouraging User Interaction of Social Network through Tweet Recommendation Using Community Structure

In this paper, we propose a tweet recommendation method that encourages people to communicate with each other on the microblogging site, "Twitter". To achieve this, we have developed a novel recommendation technique that does not only use the Bag of Words included in a tweet something written on Twitter but also human relations, i.e. followings, followers, and mentions. We also use latent Dirichlet allocation (LDA) to extract latent topics of human relations and tweets topics. Our proposal incorporates human topics in tweet topics. We present experimental results that show that the proposed method outperforms a simple tweet recommendation technique that does not use human relation information.