A Comparative Analysis of the Techniques for Influence Maximization in Social Networks

Now-a-days, social networks have become an important part of the individuals' day-to-day life. It plays a significant role in spreading data over the network. In viral advertising, organizations desire to propagate their items by utilizing the network of the organization and potentials of the influence propagation. Usually, they give the items at no cost to the chosen clients, promote them throughout the network, and maximize the acquisition. There should be a good spreading plan for the free-of-charge items with the objective of the organizations for choosing the ideal seed set to boost the influence spread. Identifying the seed nodes in the network which can maximize the influence is known as Influence Maximization (IM). Influence maximization is a principal issue to recognize a tiny arrangement of people which contain maximal impact within the community. Centrality measures help to identify the influential node in the network. This paper presents a comparative analysis of the effectiveness of the Influence Maximization techniques viz., Heuristic Algorithm (HA), Greedy Algorithm (GA), Cost Effective Lazy Forward (CELF) ++ Algorithm, and Discrete Shuffled Frog-Leaping Algorithm (DSFLA). We evaluated the performance on various social network datasets available in the public domain viz., Facebook, Skitter, Epinions, and Twitter. We observed that DSFLA is an efficient technique that outperforms in terms of the computational time as compared to the other influence maximization techniques viz., HA, GA, and CELF++.