Influence Maximization Problem Approach to Model Social Networks

Abstract

One of the significant problems in Social Net-works (SN) analysis is how to find the best influential members. This problem was proved to be non-deterministic polynomial (NP-hard). The influence maximization (IM) problem in SNs aims to maximize the spread of influence in the network. It represents an optimization problem. IM is a fundamental research problem in social networks. Influence maximization problem is the problem of assigning a subset of k users as (seed nodes) in a graph that could maximize the spread of influence by maximizing the expected number of influenced users. It represents a key algorithmic problem in social influence analysis. In this study, centrality-based methods are utilized to select top k nodes of high centrality values. The degree centrality is used to select the opinion leaders, while the between-ness and Eigenvector centrality is used to select the early adopters. The degree discount (as a heuristic approach) is proposed to replace the Greedy algorithms applied in other studies to avoid the time complexity. A mixed diffusion model (replacing the linear threshold and independent cascade) is utilized to be main diffusion model in this study.