Parallel seed selection for influence maximization based on k-shell decomposition

Abstract

Influence maximization is the problem of selecting a set of seeds in a social network to maximize the influence under certain diffusion model. Prior solutions, the greedy and its improvements are time-consuming. In this paper, we propose candidate shells influence maximization (CSIM) algorithm under heat diffusion model to select seeds in parallel. We employ CSIM algorithm (a modified algorithm of greedy) to coarsely estimate the influence spread to avoid massive estimation of heat diffusion process, thus can effectively improve the speed of selecting seeds. Moreover, we can select seeds from candidate shells in parallel. Specifically, First, we employ the k-shell decomposition method to divide a social network and generate the candidate shells. Further, we use the heat diffusion model to model the influence spread. Finally, we select seeds of candidate shells in parallel by using the CSIM algorithm. Experimental results show the effectiveness and feasibility of the proposed algorithm.