- 1、本文档共17页,可阅读全部内容。
- 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
毕业论文:SAS在聚类分析问题中的应用与研究
SAS在聚类分析问题中的应用与研究
摘 要
本文通过 SAS 软件来解决 1984 年欧洲22个国家在7个项目上的女子记录的聚类分析问题。
本文首先介绍了一些 SAS 基本知识,然后给出了与聚类分析相关的的理论知识,比如距离,相关性,并给出了具体的数学表达式。根据这些理论知识结合聚类分析的定义以及本文的研究目的,本文运用 SAS 中三种比较常用的聚类分析方法对问题进行研究分析:
一、VARCLUS 变量间聚类分析,本模型主要是对变量内的联系进行聚类分析,并给出了相关的结果:7个分量分成 5组,其中 m100 和 m200分成一组,属于短跑类型;m1500和 marathon 成为第二类,属于中长跑,而另外三个变量各成一类。
二、 FASTCLUS变量间聚类分析,本模型是对变量间进行聚类分析,得出结果如下1 类中有挪威,葡萄牙,爱尔兰,西德,英国;2 类中有西班牙,罗马尼亚,波兰,以色列,奥地利5个国家;3 类由瑞士,瑞典,荷兰,意大利,丹麦,比利时,芬兰,法国,东德。9个国家组成,4 类仅有土耳其一国;5 类由多卢森堡和希腊组成,总的来说实力由强到弱的类的顺序为 3,1,2,5,4。
三、CLUSTER 树法变量间聚类,本模型给出了22个国家在聚类过程中的具体“中间”过程,通过树的形式形象而明确的给出了分类的具体结果。
最后对三中模型的优缺点进行对比分析,本文认为各自特点鲜明,且相互补充,而且聚类结果和实际情况相吻合。
关键字:SAS 聚类分析 距离 VARCLUS FASTCLUS CLUSTER TREE
目 录
摘 要 ································ ······························1
一、研究目的 ································ ···························1
二、采用方法 ································ ···························1
三、理论知识 ································ ···························1
3.1 SAS 简介 ······························ ··························1
3.2 聚类分析定义 ······························ ······················1
3.3 聚类方法分类 ······························ ······················1
3.4 距离的相关定义 ······························ ····················2
3.5 相似系数 ······························ ··························3
3.6 类间距离定义 ······························ ······················3
3.7 聚类分析一般步骤 ······························ ·················3
四、数据的预处理 ································ ·······················4
五、具体模型 ································ ···························4
5.1 变量聚类分析 ······························ ·······················4
5.1.1 用VARCLUS过程实现变量间聚类分析 ·····························4
5.1.2 编写程序 ······························ ························4
5.1.3 输出结果 ······························ ························4
5.1.4 结果分析 ······························ ························7
5.2 FASTCLUS 变量间聚类分析 ······························ ··········7
5.2.1 用
文档评论(0)