Clustering is a frequently used unsupervised pattern recognition technique based on the grouping properties of data. K-means is one of the best known, simple and efficient method of data clustering. But this method is more sensitive to the initial cluster partitioning and suffers in local optimal cluster centers. In this paper, an attempt has been made to hybridize the K-means algorithm with the improved Particle Swarm Optimization (PSO) to improve fitness of cluster centers. The strategy of finding global best solution of IPSO is used to avoid the possibility of falling at local optimal cluster centers. The proposed method IPSO-K-means have been compared with K-means, GA-K-means and PSO-K-means and found better in terms of objective value than the others. Simulation result shows that the proposed method is effective, steady and stable and is more suitable for cluster analysis.
CITATION STYLE
Nayak, J., Naik, B., Kanungo, D. P., & Behera, H. S. (2015). An improved swarm based hybrid K-means clustering for optimal cluster centers. In Advances in Intelligent Systems and Computing (Vol. 339, pp. 545–553). Springer Verlag. https://doi.org/10.1007/978-81-322-2250-7_54
Mendeley helps you to discover research relevant for your work.