An algorithm for influence maximization in competitive social networks with unwanted users

Abstract

With the rapid development of online social networks, the problem of influence maximization (IM) has attracted tremendous attention motivated by widespread application scenarios. However, there is less research focusing on the information spreading with the existence of some unwanted users. Such problem has a wide range of applications since there always exist conflicts of interest between competing businesses. In this paper, we formally define the problem of influence maximization with limited unwanted users (IML) under the independent cascade model. In order to avoid the time-consuming process of simulating the propagation in the traditional method of influence maximization, we propose a propagation path based strategy to compute the activation probabilities between the node pairs. Based on the activation probability, we define a propagation increment function to avoid simulating the influence spreading process on the candidate seed nodes. To select the optimal seed set, we present a greedy algorithm to sequentially select the nodes which can maximize the influence increment to join the seed set. Experimental results in real social networks have shown that the algorithm proposed not only outperforms the existing methods but also consumes much less computation time.