MO-ADDOFL: Multi-objective-based adaptive dynamic directive operative fractional lion algorithm for data clustering

Abstract

Data clustering is a process that partitions the data and groups them together in groups based on their similarities and the dissimilarities existing between the data points. The concept to improve the clustering performance and to manage the imprecise and unbalanced data, this paper proposes a Multi-Objective-based Adaptive Dynamic Directive Operative Fractional Lion algorithm (MO-ADDOFL). The fitness is based on the inter-cluster distance, intra-cluster distance, and the cluster density. The fitness proposed newly in this paper enhances the quality of the clusters and improves the classification accuracy. The MO-ADDOFL operates with the maximum fitness measure and determines the optimal centroid for clustering the data points to form the optimal clusters. The proposed algorithm inherits the advantages of the fractional theory. The experimentation is performed using two benchmark datasets, such as lung cancer and Pima Indians Diabetes dataset, which proves that the classification accuracy of the proposed method is better compared with other clustering algorithms. The maximum classification accuracy of MO-ADDOFL is found as 0.84, for the Pima Indians Diabetes dataset.