摘要
给出了选用九种相似性度量,用最短距离法聚类,结果互不相同的一个有趣的例子。对该例,用欧氏距离求出距离矩阵后,除用最短距离法聚类结果唯一外,用最长距离法、重心法、类平均法、离差平方和法聚类,结果均不唯一。
In this paper, we give an interesting example. First,using nine similarity measures, we obtain distance matrices. and then, applying nearest -neighbour method, the cluster results are different each other, Second, using Euclidean distance,we obtain a distance matrix .Then, applying furthestneighbour method, centroid method,group-average method and sum of squares method, exceptnearest-neighbour method every method, can obtain two cluster esurIts.
出处
《数理统计与管理》
CSSCI
北大核心
1993年第6期55-59,共5页
Journal of Applied Statistics and Management