WGC: Hybridization of exponential grey wolf optimizer with whale optimization for data clustering

Abstract

Data present in abundance increases the complexity of handling them, which affects the effective decision-making process. Hence, data clustering gains remarkable importance in knowledge extraction and an efficient clustering method promotes the efficient decision making. Therefore, this paper proposes a method of data clustering using the WGC algorithm that determines the optimal centroid for performing the clustering process. The WGC algorithm uses the computation steps of the Whale Optimization Algorithm (WOA) with the integration of the WEGWO with a newly formulated fitness function. The WGC algorithm uses the minimum fitness measure to locate the optimal centroid and the fitness measure depends on three constraints, namely inter-cluster distance, intra-cluster distance, and the cluster density. The optimal centroid corresponding to the minimal value of the fitness is utilized for clustering the data. Experimentation is carried out using three datasets and the comparative performance analysis is carried out, which proves that the proposed WGC outperforms the existing methods. The proposed WGC attained a maximum value of F-measure, rand coefficient, and Jaccord coefficient at a rate of 0.9716, 0.9695, and 0.8949 and a minimum value of 1.463 for MSE proving that the proposed WGC algorithm outperformed the existing Particle Swarm Clustering (PSC), modified Particle Swarm Clustering (mPSC), Grey Wolf Optimization (GWO), Exponential Grey Wolf Optimization (EGWO), Kernel-based exponential grey wolf optimizer (KEGWO), and WOA.