A novel chemical reaction-based clustering and its performance analysis. (1st January 2013)

• Record Type: Journal Article
• Authors: Baral, Archana; Behera, H.S.
• Is Part Of: International journal of business intelligence and data mining. Volume 8:Number 2(2013)
• Page Start: 184
• Page End: 198
• Publication Date: 2013-01-01
• Subjects: clustering -- K-means -- genetic algorithm -- differential evolution -- teaching learning-based optimisation -- chemical reaction-based optimisation -- clustering algorithms; 006.312
• Languages: English
• ISSNs: 1743-8187

Abstract

Clustering is widely used in case of pattern recognition, decision-making, machine-learning, image processing and many more real world problems. Many algorithms have been developed for better clustering. This paper proposes an efficient way of clustering data using chemical reaction optimisation (CRO), a recently developed metaheuristics for solving optimisation problems. By taking into consideration some of the real world datasets, the performance of the proposed algorithm has been compared with K-means, genetic algorithm (GA), differential evolution (DE) and teaching-learning-based optimisation (TLBO). Experimental result shows that the performance of CRO-based clustering is better than K-means, GA, DE and TLBO, in terms of quantisation error, intra and inter cluster distance, etc. It is also observed that the proposed CRO-clustering algorithm converges remarkably faster in comparison to other algorithms.