半径自适应的初始中心点选择K-medoids聚类算法  

Radius-Adaptive on Selection of Initial Centers in K-Medoids Clustering

作　　者：王勇[1] 王李福 饶勤菲 邹辉[1] WANG Yong WANG Li-fu RAO Qin-fei ZOU Hui(College of Computer Science and Engineering, Chongqing University of Technology, Chongqing 400054, Chin)

机构地区：[1]重庆理工大学计算机科学与工程学院,重庆400054

出　　处：《重庆理工大学学报（自然科学）》2017年第2期95-101,共7页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(61173184)

摘　　要：针对K-medoids(K为中心点)聚类算法对初始聚类中心敏感、聚类结果依赖于初始聚类中心的缺陷,提出一种新的半径自适应的初始中心点选择算法。该算法在每次迭代过程中都重新根据剩余样本点的分布特征计算半径,从而实现动态计算相应样本点的局部方差和领域半径,选取较优的初始聚类中心点,实现良好的聚类效果。采用不同规模的UCI数据集和不同比例随机点的模拟数据集进行测试,利用5个通用的聚类评价指标对性能进行评价。结果表明:本算法性能较同类算法有明显提高。For the defect of K-medoids clustering algorithm which was sensible to the initial center and depended on the initial center,this paper proposed a new selection of initial centers algorithm through radius adaptive.The algorithm calculated each radius according to distribution of the remaining sample in each iteration,thus to dynamically calculate local variance and neighborhood radius for corresponding sample and then select the optimum initial cluster centers to achieve good clustering effect.Using different size of UCI data sets and simulated data sets with different ratios of random points for testing and using the five general clustering index to evaluate its performance,the results show that the performance of this algorithm is better than other similar algorithms.

关 键 词：局部方差 初始聚类中心 聚类 K-medoids 自适应 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]