|
|||||
|
|
| 一种基于熵和均方差法综合赋权的K-means算法 | |
| 引用本文： | 上官廷华,冯荣耀,柳宏川.一种基于熵和均方差法综合赋权的K-means算法[J].计算机与现代化,2010(4):34-36. |
| 作者姓名： | 上官廷华 冯荣耀 柳宏川 |
| 作者单位： | 1. 河南商业高等专科学校现代教育与实验中心,河南,郑州450045 2. 郑州大学教育技术中心,河南,郑州,450052 3. 郑州大学信息工程学院,河南,郑州,450052 |
| 摘 要： | 在传统的K-means聚类算法基础上,本文提出一种基于熵和均方差法综合赋权的Syn-K-means算法。引入综合权重提高聚类结果的类内相似度,从而提高聚类精度。算法中特征权重的计算基于概率论中数字特征的基本描述方法——均方差和信息论中信息特征的基本度量方法——熵;综合赋权系数的选择采用主观设定法求解。实验结果表明,Syn-K-means算法在聚类精度方面优于标准的K-means算法。 |
| 关 键 词： | K-均值算法 综合权重 均方差 熵 |
An Attribute Weighting K-means Algorithm Based on Synthetic Weight |
|
| SHANGGUAN Ting-hua,FENG Rong-yao,LIU Hong-chuan.An Attribute Weighting K-means Algorithm Based on Synthetic Weight[J].Computer and Modernization,2010(4):34-36. | |
| Authors: | SHANGGUAN Ting-hua FENG Rong-yao LIU Hong-chuan |
| Affiliation: | 1.Center of Education and Experiment/a>;Henan Business College/a>;Zhengzhou 450045/a>;China/a>;2.Center of Education Technical/a>;Zhengzhou University/a>;Zhengzhou 450052/a>;3.College of Information Engineering/a>;China |
| Abstract: | This paper proposes an attribute weighting K-means algorithm based on synthetic weight in order to improve the clustering quality.The weighting schemes are Entropy weighting scheme and Mean-square-deviation weighting scheme.Experimental results on real life datasets show that it is superior to algorithms without weight with respect to adjusted rand index measures. |
| Keywords: | K-means synthetic weight mean-square-deviation entropy |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |
