A novel hybrid approach for real world data clustering algorithm based on fuzzy C-means and firefly algorithm. (2015)

• Record Type: Journal Article
• Authors: Behera, Himansu Sekhar; Nayak, Janmenjoy; Nanda, M.; Nayak, K.
• Is Part Of: International journal of fuzzy computation and modelling. Volume 1:Number 4(2015)
• Page Start: 431
• Page End: 448
• Publication Date: 2015
• Subjects: particle swarm optimisation -- PSO -- fuzzy C-means -- FCM -- firefly algorithm -- data clustering -- fuzzy clustering; Fuzzy mathematics -- Periodicals; Mathematical models -- Periodicals; Computer science -- Mathematics -- Periodicals; 004.01511313
• Languages: English
• ISSNs: 2052-353X

Abstract

Fuzzy clustering plays an important role in the current research area for several real world applications. But as it is highly sensitive to initialisations and suffers from local optima, it is not being well suited for many real world problems. Firefly algorithm is a nature inspired metaheuristic optimisation algorithm which is inspired by the simulation of flashing behaviour of fireflies. In order to overcome the shortcomings of fuzzy C-means (FCM) algorithm, a hybridised fuzzy clustering algorithm has been proposed by combining the fuzzy C-means with firefly (FCM-FA) algorithm by extracting the positive insights of both the algorithm for faster convergence than FCM. The performance of proposed hybridised FCM-FA has been compared with other used algorithms for various real world datasets. The experimental results show that the hybrid FCM-FA is better and more effective clustering method for real world applications as compared to other existing algorithms.