Approximation Methods for Influence Maximization in Temporal Networks

Abstract

The process of rumor spreading among people can be represented as information diffusion in a social network. The scale of the rumor spread can change greatly depending on the starting nodes. If we can select nodes that trigger large scale diffusion events, the nodes are expected to be important for viral marketing. Given a network and the number of starting nodes, the problem of selecting nodes for maximizing information diffusion is called as influence maximization problem. We propose three new approximation methods (Dynamic Degree Discount, Dynamic CI, and Dynamic RIS) for influence maximization problem in temporal networks. These methods are the extensions of previous methods for static networks to temporal networks. Although the performance of MC greedy was better than the three methods, it was computationally expensive and intractable for large-scale networks. The computational time of our proposed methods was more than 10 times faster than MC greedy. When compared with Osawa, the performances of the three methods were better for most of the cases.