Opinion Maximization in Signed Social Networks Using Centrality Measures and Clustering Techniques

In a social network, Opinion Maximization is a problem that targets spreading the desired opinion across the social network. In the real world, every user/individual has their own opinion. The opinion of an individual can be favorable or unfavorable. Each individual can affect the others positively or negatively. In this paper, we consider the Opinion Maximization problem for signed, weighted and directed social networks with Multi-Stage Linear Threshold Model for information propagation. The seed nodes are responsible for spreading the desired opinion in the network. The Opinion Maximization problem asks to compute the minimum number of seed nodes such that the overall opinion of the network is maximized. In this paper, we proposed three heuristics to select the seed nodes. The proposed methods use centrality measures and clustering of the social network. The proposed methods are tested on real-world as well as synthetic data sets.