Chapter 9 Cluster analysis
zf Chapter 9 Cluster analysis
Presentation outline(本章要点) .o What is cluster analysis? Similarities measures .g Hierarchical cluster analysis Centroid method Single linkage Complete linkage Average linkage Ward s method Number of clusters Non-hierarchical cluster analysis 2021/2/22 2 cxt
2021/2/22 2 cxt Presentation Outline(本章要点) ❖ What is cluster analysis? ❖ Similarities measures ❖ Hierarchical cluster analysis – Centroid method – Single linkage – Complete linkage – Average linkage – Ward’s method – Number of clusters ❖ Non-hierarchical cluster analysis
一、什么是聚类分析 What is cluster analysis? 令1、 definition(定义) Cluster analysis is a technique used for combining observations into groups or clusters such that (1) Each group or cluster is homogeneous or compact with respect to certain characteristics. That is observations in each group are similar to each other (2) Each group should be different from other groups with respect to the same characteristics that is observations of one group should be different from the observations of other groups 2021/2/22 cxt
2021/2/22 3 cxt 一、什么是聚类分析What is cluster analysis? ❖ 1、definition(定义) Cluster analysis is a technique used for combining observationsinto groups or clusters such that: (1) Each group or cluster is homogeneous or compact with respect to certain characteristics. That is, observations in each group are similar to each other. (2) Each group should be different from other groups with respect to the same characteristics; that is, observations of one group should be different from the observations of other groups
◆聚类分析 是根据“物以类聚”的道理,对样品或指标 进行分类的一种多元统计分析方法。 将个体或对象分类,使得同一类中的对象之 问的相似性比与其他类的对象的相似性更强。 聚类分析的目的 使类内对象的同质性最大化和类间对象的 异质性最大化。 2021/2/22 4 cxt
2021/2/22 4 cxt ❖ 聚类分析 是根据“物以类聚”的道理,对样品或指标 进行分类的一种多元统计分析方法。 将个体或对象分类,使得同一类中的对象之 间的相似性比与其他类的对象的相似性更强。 ❖ 聚类分析的目的 使类内对象的同质性最大化和类间对象的 异质性最大化
◆聚类分析的基本思想: 是根据一批样品的多个观测指标,具体地找出 些能够度量样品或指标之间相似程度的统计 量,然后利用统计量将样品或指标进行归类。 把相似的样品或指标归为一类,把不相似的 归为其他类。直到把所有的样品(或指标) 聚合完毕. ◇相似样本或指标的集合称为类。 2021/2/22 5 cxt
2021/2/22 5 cxt ❖ 聚类分析的基本思想: 是根据一批样品的多个观测指标,具体地找出 一些能够度量样品或指标之间相似程度的统计 量,然后利用统计量将样品或指标进行归类。 把相似的样品或指标归为一类,把不相似的 归为其他类。直到把所有的样品(或指标) 聚合完毕. ❖ 相似样本或指标的集合称为类