Hybrid K-means Algorithm and Genetic Algorithm for Cluster Analysis

Dianhu Cheng (Ocean University of China)
Xiangqian Ding (Ocean University of China)
Jianxin Zeng (China Tobacco Yunnan Industrial Co., Ltd)
Ning Yang (Qingdao Ocean University)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 12, No 4: April 2014

Cluster analysis isa fundamental technique for various filed such as pattern recognition, machinelearning and so forth. However, the cluster number is predefined by users inK-means algorithm, which is unpractical to implement.  Since the number of clusters is a NP-completeproblem, Genetic Algorithm is employed to solve it. In addition, due to the largetime consuming in conventional method, an improved fitness function isproposed. According to the simulation results, the proposed approach isfeasible and effective. DOI : http://dx.doi.org/10.11591/telkomnika.v12i4.4805