The research of K-medoids clustering algorithm based on density
Ping Liu1, Hao Zhou1, Junping Yang2, Taorong Qiu1
COMPUTER MODELLING & NEW TECHNOLOGIES 2013 17(5A) 7-11
1 The school of Information Engineering, Nanchang University, Nanchang 330031, China
2 Affiliated Hospital of Jiangxi College of Traditional Chinese Medicine, Nanchang 330006, China
In view of that the clustering result of the traditional k-medoids clustering algorithm being sensitive to initial cluster centers. A new kmedoids clustering algorithm based on density was proposed in this paper. It conducted a rough clustering to generate several particles at first. Then select the centers of the k densest particles as the initial clustering centers. Tested by using UCI data sets, the validity of the proposed algorithm is demonstrated.