Influence Maximization Using User Connectivity Guarantee in Social Networks

Abstract

With the rapid development of social networks, the influence maximization problem has attracted more and more attention from academia and industry. Its aim is to find a set of nodes as seeds to spread the influence as widely as possible. However, most of the existing researches neglected the connectivity of seeds, which has effect on the process of information diffusion. In this paper, we propose a novel problem, connectivity guaranteed influence maximization, which suggests a fixed number of new links to the seed set with the aim of maximizing the influence of seed nodes while guaranteeing the connectivity of the induced subgraphs consisting of active nodes. To tackle this problem, we propose a Connectivity Guaranteed Influence Maximization (CGIM) algorithm based on user connec-tivity and link recommendation. Specifically, Jaccard coefficient is first used to calculate the influence between users. Then a Connectivity Guarantee based Link Addition (CGLA) algorithm is proposed to keep the connectivity of the induced sub graphs formed by all active nodes after influence propagation. Following that, an improved approximate influence maximization algorithm is proposed to maximize the influence by recommending a number of new links to the seed set. Experimental results on real social network datasets show that the proposed CGIM algorithm can maximize the influence of seed nodes while guarantee user connectivity. and has good performance and scalability.