Influence maximization in social networks using community-diversified seed selection

Abstract

The community-based influence maximization in social networks refers to the strategy of amplifying the influence by leveraging the underlying connectivity patterns of the online communities. This promotes the widespread dissemination of user-generated content across socially connected entities. In our work, we have combined the simplicity of degree centrality and the bridging strength of betweenness for maximizing the influence performance. We have used community detection for partitioning the whole network into smaller subnetworks for intelligent distribution of seed nodes among these communities for improving the rate of information spreading. The proposed ranking, named as Community Diversified Seed Selection, is compared with degree centrality and betweenness centrality-based ranking in terms of rate of spreading, absolute execution time and algorithmic time complexity. The comparative study was performed on LFR${}_{\mu }=$ 0.01, LFR${}_{\mu }=$ 0.02 and LFR${}_{\mu }=$ 0.03 benchmark networks, and validated using Facebook real-world social network. Our proposed algorithm has better and faster spreading with reasonable time complexity, therefore, making it a suitable choice for larger networks.