An Influence Model Based on Heterogeneous Online Social Network for Influence Maximization

Influence maximization is an important technique in advertisement post, viral marketing, and public opinion monitoring. Seed set identification is one of the key issues in influence maximization. In reality, there exist heterogeneous nodes, such as user nodes, message nodes in social networks. The complex association relationship among heterogeneous nodes, which are seldom considered, significantly increases the complexity of the seed set identification. In this paper, we propose a Measuring Influence (MIF) model to capture social influence with heterogeneity. MIF considers the interaction among adjacent nodes, the tag of users, the users' social friendships and the similarity in user interests, and studies the interaction based influence, tag based influence, friendship based influence, and topic based influence, respectively. As obtaining the seed set in social networks has been proved to be a NP-hard problem, we propose an algorithm called Influence Maximization Greedy Algorithm (IMGA) to solve this problem by maximizing the marginal influence of selected seed nodes. Series of experiments are designed to evaluate the performance of the proposed model and algorithm. Our results show that MIF model and IMGA algorithm have better influence spread effects and higher quality of the seed set identification comparing to the approaches under IC, LT, CDNF, MIA, and BBA, models.