Title: A modified differential evolution-based fuzzy multi-objective approach for clustering

Authors: Subrat Kumar Nayak; Pravat Kumar Rout; Alok Kumar Jagadev; Srikanta Patnaik

Addresses: Department of Computer Science, Siksha 'O' Anusandhan University, Bhubaneswar-30,Odisha, India ' Department of Electrical and Electronics Engineering, Siksha 'O' Anusandhan University, Bhubaneswar-30, Odisha, India ' School of Computer Engineering, KIIT University, Bhubaneswar, Odisha, India ' Department of Computer Science, Siksha 'O' Anusandhan University, Bhubaneswar-30,Odisha, India

Abstract: Many evolutionary-based metaheuristics have been proposed for minimisation of intra cluster distance for better clustering, however the effect of inter cluster distances on clustering is ignored in most of the cases. Considering this issue a modified differential evolution-based fuzzy multi-objective (MDEFM) approach is proposed in this study where effect of both intra and inter cluster distance on clustering is analysed. This way of considering both the objectives and assigning different weighting factors according to their priority results in well separate clusters with greater accuracy. Furthermore, a centroid rearrangement scheme has been proposed for getting a consistent result. A comparative analysis of the proposed approach with another five population-based methods on eight real datasets is carried out to justify the efficacy of the model. The results reveal that the proposed approach can be considered as one of the alternate powerful methods for data clustering applications in various fields.

Keywords: differential evolution; data clustering; inter cluster distance; intra cluster distance; centroid rearrangement; fuzzy clustering; multi-objective clustering; metaheuristics.

DOI: 10.1504/IJMDM.2017.082509

International Journal of Management and Decision Making, 2017 Vol.16 No.1, pp.24 - 49

Available online: 23 Feb 2017 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article