桂林电子科技大学学报2013，Vol.33Issue(2)：134-138,5.
一种基于半径的k-means算法
A k-means algorithm based on the radius
摘要
Abstract
The initial center of the clusters in the traditional k-means algorithm and its improved algorithms k-means + + is random,Clustering results are likely to fall into the local minima,which makes its application in some practical problems invalid.In order to overcome this drawback,a k-means+λ algorithm based on the radius is proposed,the initial cluster center is selected by calculating the distance between each point according to the parameter λ and distance ratio in the circle with a certain radius.In the same test environment,KDD CUP99 data set is used.The experimental results show that compared to the k-means algorithm and k-means+ + algorithm,this algorithm is effective on the error rate and computation time.关键词
k-means/聚类/k-means++/k-means+λ/距离Key words
k-means/ cluster/ k-means++/ k-means+λ distance分类
信息技术与安全科学引用本文复制引用
刘家星,朱国魂,席敏..一种基于半径的k-means算法[J].桂林电子科技大学学报,2013,33(2):134-138,5.基金项目
广西科学研究与技术开发计划(1114006-3C) （1114006-3C）
