| 基于层次与划分方法的聚类算法研究 |
| |
| 引用本文： | 甄彤. 基于层次与划分方法的聚类算法研究[J]. 计算机工程与应用, 2006, 42(8): 178-180 |
| |
| 作者姓名： | 甄彤 |
| |
| 作者单位： | 华中科技大学控制科学与工程系,武汉,430074;河南工业大学信息科学与工程学院,郑州,450052 |
| |
| 摘 要： | 针对在层次聚类算法中,一个分裂或合并被执行,就不能修正,其聚类质量受到限制的缺陷,提出了利用簇间相异度及基于信息熵或整体相似度的聚类质量评价标准,在簇分裂过程中动态的进行簇的合并与分裂的算法。仿真实验结果证明,该算法具有使结果簇更紧凑和独立的效果,具有更好的聚类质量。
|
| 关 键 词： | 层次聚类 相异度 信息熵 整体相似度 聚类质量 |
| 文章编号： | 1002-8331-（2006）08-0178-03 |
| 收稿时间： | 2005-12-01 |
| 修稿时间： | 2005-12-01 |
Research of Clustering Algorithm Based on Hierarchical and Partitioning Method |
| |
| Zhen Tong. Research of Clustering Algorithm Based on Hierarchical and Partitioning Method[J]. Computer Engineering and Applications, 2006, 42(8): 178-180 |
| |
| Authors: | Zhen Tong |
| |
| Affiliation: | Department of Control Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074 |
| |
| Abstract: | Hierarchical methods suffer from the fact that once a merge or split is done,it can never be undone.This paper presents a clustering algorithm based on hierarchical and partitioning method.During the split process,clusters are merged and split dynamically by using dissimilarity measure between clusters and entropy or overall similarity to evaluate the cluster quality.The experiment shows the better results with more compactness and separation. |
| |
| Keywords: | hierarchical clustering dissimilarity measure entropy overall similarity cluster quality |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
