Performance Analysis of Whale optimization based Data Clustering

Abstract

Data clustering is the method of gathering of data points so that the more similar points will be in the same group. It is a key role in exploratory data mining and a popular technique used in many fields to analyze statistical data. Quality clusters are the key requirement of the cluster analysis result. There will be tradeoffs between the speed of the clustering algorithm and the quality of clusters it produces. Both the quality and speed criteria must be considered for the state-of-the-art clustering algorithm for applications. The Bio-inspired technique has ensured that the process is not trapped in local minima, which is the main bottleneck of the traditional clustering algorithm. The results produced by the bio-inspired clustering algorithms are better than the traditional clustering algorithms. The newly introduced Whale optimization-based clustering is one of the promising algorithms from the bio-inspired family. The quality of clusters produced by Whale optimization-based clustering is compared with k-means, Kohonen self-organizing feature diagram, Grey wolf optimization. Popular quality measures such as the Silhouette index, Davies-Bouldin index, and Calianski-Harabasz index are used in the evaluation.