Multi-objective clustering: a kernel based approach using Differential Evolution. Issue 3 (3rd July 2019)

• Record Type: Journal Article
• Authors: Nayak, Subrat Kumar; Rout, Pravat Kumar; Jagadev, Alok Kumar
• Is Part Of: Connection science. Volume 31:Issue 3(2019)
• Page Start: 294
• Page End: 321
• Publication Date: 2019-07-03
• Subjects: Multi-objective clustering -- Differential Evolution -- kernel clustering -- Pareto front; Neural computers -- Periodicals; Artificial intelligence -- Periodicals; Cognitive science -- Periodicals; Connectionism -- Periodicals; 006.3
• DOI: 10.1080/09540091.2019.1603201
• Languages: English
• ISSNs: 0954-0091

Abstract

ABSTRACT: A multi-objective algorithm is always favoured over a single objective algorithm as it considers different aspects of a dataset in the form of various objectives. In this article, a multi-objective clustering algorithm has been proposed based on Differential Evolution. Here, three objectives have been considered to handle different complex datasets. In addition to this, a kernel function is hybridised with the objectives to evaluate the data on a hyperspace for reducing the impact of nonlinearity on cluster formation. Moreover, to get the best compromised solution from the Pareto front an effective fuzzy concept has been followed. Five metaheuristic approaches have been taken into consideration for performance comparison. These methodologies have been applied to twelve datasets and the result reveals the efficacy of the proposed model in data clustering.