Cost Optimized Community-Based Influence Maximization

Effective identification of a small set of nodes within a network which can potentially cover many number of nodes in the remaining network is known as influence spread process. Influence spreading amongst maximum number of nodes is called influence maximization process. Influence maximization task is computationally hard which involves promising seed set selection and estimation of the maximum influence spread throughout the network. Community detection algorithm to figure out effective seed set for influence maximization within an acceptable execution time is the key essence of this article. Proposed community-based identification method involves three stages. First stage detects communities in the given network, second stage analyzes community structure to select candidate nodes within the communities, and the third stage identifies promising influential members from the candidate set to make a target set. Ultimately, average influence spread is measured through Monte Carlo simulation technique. Proposed algorithm has been rigorously tested on two real-world social network datasets to establish its usefulness and efficacy.