| 一种改进的K-means聚类算法 |
| |
| 引用本文： | 张建民.一种改进的K-means聚类算法[J].微计算机信息,2010(9). |
| |
| 作者姓名： | 张建民 |
| |
| 作者单位： | 安徽省地方税务局信息中心; |
| |
| 摘 要： | 本文提出了一种带离群点数据过滤的K-means改进算法。该算法根据离群点数据特征制定了离群点数据的发现规则,并在原算法中加入了离群点数据的发现和处理步骤。通过对给定的具有普遍意义的数据实验表明,改进后算法能较为稳定的发现数据集中存在的离群点数据,这些离群点数据符合离群点数据特征;同时在剔除这些极少数离群点数据后,显著提高了聚类结果簇的凝聚度,从而有效克服了离群点数据对K-means算法的影响,使聚类效果得以显著提高。
|
| 关 键 词： | 聚类分析 K-means算法 离群点数据 |
An Improved K-means Clustering Algorithm |
| |
| ZHANG Jian-min.An Improved K-means Clustering Algorithm[J].Control & Automation,2010(9). |
| |
| Authors: | ZHANG Jian-min |
| |
| Abstract: | An improved K-means clustering algorithm based on outlier data filtering is presented in this paper.This algorithm makesfinding rule of outlier data according to its data characteristic and adds the procedure of finding and disposing outlier data in the o-riginal one.In the experiment which uses the given data,the result shows that this improved algorithm can find the outlier datasteadily;after eliminating these outlier data,the cohesion of result cluster is improved markedly,so,this improved algorithm can ... |
| |
| Keywords: | Clustering Analysis K-means Algorithm outlier data |
| 本文献已被 CNKI 等数据库收录！ |
|
