Effective Method for Promoting Viral Marketing in Microblog

Based on word-of-mouth effect, viral marketing has developed into an important marketing strategy recently. Meanwhile, microblog has drawn global attention as a potential marketing group over the past decade. Thus how to promote viral marketing in microblog has become a hot topic in social network. However, finding the most influential users for viral marketing under some diffusion models has been proven to be NP-hard [2]. Even though several algorithms have been proposed to approximate this problem, some of them are time-consuming, some inaccurate, and some including too many assumptions. In this paper, we propose an effective method which can obtain a good approximation with nice speed. To reduce the number of candidate users, we present MBRank to rank all the users in microblog and select a small number of candidates. Based on the small candidate set, MBGreedy and its improvement MBCELF for influence maximization are given. They combine both the advantages of greedy-based and heuristic-based algorithms. Furthermore, whether the top-k ranked users always lead to influence maximization is involved. Extensive experiments are conducted to prove that MBCELF is a good tradeoff between effectiveness and efficiency, and the experimental results reveal that our method is really competent in promoting viral marketing in microblog.