A Learning Automaton-Based Algorithm for Influence Maximization in Social Networks

Jinchao Huang,Hao Ge,Ying Guo,Yan Zhang,Shenghong Li
DOI: https://doi.org/10.1007/978-981-10-3229-5_76
2017-01-01
Abstract:Influence maximization problem is to find a small set of influential nodes in a social network, whose activation of information propagation can be maximized using one of the propagation models such as independent cascade model. This paper proposes a new method to solve this problem based on discretized linear automaton algorithm and simple greedy algorithm. Every allowable node in the social network is regarded as an action and the goal of the learning automaton is to select an optimal subset of actions. To speed up the convergence, firstly, the problem space is reduced by excluding the nodes lacking influence; then, the original learning process is divided into two parts: the sampling process and the learning process. Besides, a new scheme for determining the best parameters is presented, which is more applicable to practical problems. The obtained results show that the new proposed algorithm is efficient in real-life social networks. The results are close to the ones obtained by the greedy algorithm in terms of accuracy, and the new method greatly improves the speed of selecting the optimal subset.
What problem does this paper attempt to address?