A Hybrid Data Clustering Approach Based on Cat Swarm Optimization and K- Harmonic Mean Algorithm
Year: 2014
Journal of Information and Computing Science, Vol. 9 (2014), Iss. 3 : pp. 196–209
Abstract
Clustering is an important task that is used to find subsets of similar objects from a set of objects such that the objects in the same subsets are more similar than other subsets. Large number of algorithms has been developed to solve the clustering problem. K-Harmonic Mean (KHM) is one of the popular technique that has been applied in clustering as a substitute of K-Means algorithm because it is insensitive to initialization issues due to built in boosting function. But, this method is also trapped in local optima. On the other hand, Cat Swarm Optimization (CSO) is the latest population based optimization method used for global optimization. In this paper a hybrid data clustering method is proposed based on CSO and KHM which includes the advantage of both algorithms and named as CSOKHM. The hybrid CSOKHM not only improved the convergence speed of CSO but also escape the KHM method to run in local optima. The performance of the CSOKHM is evaluated using seven datasets and compared with KHM, PSO, PSOKHM, ACA, ACAKHM, GSAKHM, CSO methods. The experimental results show the applicability of CSOKHM method..
Journal Article Details
Publisher Name: Global Science Press
Language: English
DOI: https://doi.org/2024-JICS-22580
Journal of Information and Computing Science, Vol. 9 (2014), Iss. 3 : pp. 196–209
Published online: 2014-01
AMS Subject Headings:
Copyright: COPYRIGHT: © Global Science Press
Pages: 14