Opinion leader detection using whale optimization algorithm in online social network

In the current digital era, optimization is one of the most significant problems in the social network. Most of the issues related to optimization are NP-complete and not possible to solve them in polynomial time. Detection of opinion leader based on their optimized centrality measure is a critical issue. The opinion leaders have a non-trivial influence on the other user's decision-making process and can solve various problems related to the diffusion of new products and innovations in the real world. In this paper, we proposed a new Social Network-based Whale Optimization Algorithm (SNWOA) to find the top-N opinion leaders by measuring the reputation of the user using various standard optimization function in the network. The proposed algorithm is advantageous to determine the opinion leaders because it based on the bubble-net hunting behavior of humpback whales. The algorithm found the best possible solution as the number of users raises progressively in the network; therefore, the general complexity of the algorithm remains unchanged. Besides, we also proposed a new approach to categorize the communities based on the similarity index comprising neighbor similarity and clustering coefficient as the significant components. Initially, we computed the objective function of each user by using their centralities and deployed the proposed algorithm with different optimization functions to identify the local and universal opinion leaders. We implemented the proposed algorithm on the real and synthesized datasets and compared the result based on the accuracy, precision, recall, and F1-score. The result indicates that the proposed algorithms give a better result as compared to the other standard Social Network Analysis (SNA) measures. We also concluded that the community partitioning algorithm is even better than the other community detection algorithms based on different parameters and computational time.