学报首页    
学报动态        
· 重要声明:长春理工大学学报投稿邮箱为custlxb@cust.edu.cn
更多>>
投稿指南        
相关下载        
· 文章模板
· 版面费办理办法
· 保密审查表
更多>>
友情链接        
· 长春理工大学
· 长春理工大学图书馆
· 中国知网
当前位置:首页»自然学科» 当期目录

基于加权处罚的K-均值优化算法

发布日期:2014-07-08| 阅读次数: | 关键字:38-4 | 作者:梁鲜,曲福恒,杨勇,才华 | 来源:长春理工大学学报:自然科学版 2015 Vol.38(4): 132-137

基于加权处罚的K-均值优化算法

梁鲜1,曲福恒1,杨勇1,才华2

1.长春理工大学计算机科学技术学院,长春1300222.长春理工大学电子信息工程学院,长春130022

摘要:在各种聚类算法中,基于目标函数的K-均值聚类算法应用最为广泛,然而,K-均值算法对初始聚类中心特别敏感,聚类结果易收敛于局部最优。为此,提出基于加权处罚的K-均值优化算法。每次迭代过程中,根据簇的平均误差的大小为簇分配权值,构造加权准则函数,把样本分给加权距离最小的簇中。限制簇集中出现平均误差较大的簇,提高聚类准确率。实验结果表明,该算法与K-均值算法、优化初始聚类中心的K-均值算法相比,在含有噪音的数据集中,表现出更好的抗噪性能,聚类效果更好。

关键词:聚类;K-均值算法;初始聚类中心;聚类准则函数

中图分类号: TP391.9 文献标识码:A 文章编号:1672-9870201503-0132-06

 

An Optimal K-means Algorithm Based on Weighted Penalty

LIANG Xian1QU Fuheng1YANG Yong1CAI Hua2

1. School of Computer Science and TechnologyChangchun University of Science and TechnologyChangchun 130022

2. School of Electronics and Information EngineeringChangchun University of Science and TechnologyChangchun 130022

AbstractAbstractIn a variety of clustering algorithmsK-means clustering algorithm which is based on the objective function has the most widely usedHoweverK-means is sensible to the initial seedspoor local optima can be easily obtained. To tackle the initialization problem of K-meansan optimal K-Means algorithm based on weighted penalty is proposed. In each iteration process the weights are assigned for the clusters relative to their average variance a weighted version of K-means objective is constructedthe samples are taken to the clusters of minimum weighted distance. The emergence of large average variance clusters is limited and the clustering accuracy is improved. The effectiveness of the approach is verified in experiments and the immune property with noises is got in its clusteringas it is compared favorably with both K-means and other methods from the literature that consider the K-means initialization problem.

Key wordsclusteringK-means algorithminitial clustering centerclustering criterion function

 

基金项目:吉林省自然科学基金(201215145);吉林省自然科学基金(20130101179JC-13);吉林省教育厅科研项目(2013-420

作者简介:梁鲜(1990-),女,硕士研究生,E-mailliangxian1143@163.com

通讯作者:曲福恒(1976-),男,博士,副教授,E-mail1179954525@qq.com

版权所有:长春理工大学学报编辑部
Copyright ©Changchun University of Science and Technology