GA with k-Medoid Approach for Optimal Seed Selection to Maximize Social Influence

In this rapidly rising field of Web, volume of online social networks has increased exponentially. This inspires the researchers to work in the area of information diffusion, i.e., spread of information through “word of mouth” effect. Information maximization is an important research problem of information diffusion, i.e., selection of k most influential nodes in the network such that they can maximize the information spread. In this paper, we proposed an influence maximization model that identifies optimal seeds to maximize the influence spread in the network. Our proposed algorithm is a hybrid approach, i.e., GA with k-medoid approach using dynamic edge strength. To analyze the efficiency of the proposed algorithm, experiments are performed on two large-scale datasets using fitness score measure. Experimental outcome illustrated 8–16% increment in influence propagation by proposed algorithm as compared to existing seed selection methods, i.e., general greedy, random, discounted degree, and high degree.