| 聚类分析计算方法的理论及结果比较 |
| |
| 引用本文： | 吴元奇,冯荣扬.聚类分析计算方法的理论及结果比较[J].广东海洋大学学报,2002,22(1):57-63. |
| |
| 作者姓名： | 吴元奇 冯荣扬 |
| |
| 作者单位： | 湛江海洋大学农学院,广东,湛江,524088 |
| |
| 摘 要： | 通过对聚类分析中几种数据规格化方法、相似系数计算方法以及聚类递推公式进行分析 ,并以雷州半岛芒果炭疽病发生的生态区划为实例进行研究。结果表明 :均值规格化是聚类分析中最好的规格化方法 ;相似系数可分为绝对差异相似性、相对差异相似性和比例相似性三类 ;各种递推公式所得的聚类结果树谱图及其拓朴矩阵表明最短距离法的聚类结果与模糊聚类结果相同
|
| 关 键 词： | 聚类分析 计算理论 生态区划 芒果炭疽病 |
| 文章编号： | 1007-7995(2002)01-0057-07 |
| 修稿时间： | 2001年9月23日 |
Studies on Algorithmic Theories and Result Comparisons of Cluster Analysis |
| |
| Wu Yuanqi,Feng Rongyang.Studies on Algorithmic Theories and Result Comparisons of Cluster Analysis[J].Journal of Zhanjiang Ocean University,2002,22(1):57-63. |
| |
| Authors: | Wu Yuanqi Feng Rongyang |
| |
| Abstract: | A few methods of treating data, calculating similarity coefficients, and cluster inferential formula used in cluster analysis were compared, and the ecological planning of mango anthracnose in the Leizhou Peninsula was studied as a case. The results showed that, standardizing data with being divided by the arithmetic average is the best among all methods; the similarity coefficients can be classified into three kinds: the absolute difference similarities, the relative difference similarities, and the ratio similarities; the dendrogram and its topological structure indicate that the results of the least distance clustering are the same as that of fuzzy cluster analysis. |
| |
| Keywords: | Cluster analysis Algorithmic theory Ecological planning Mango anthracnose |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
