合理进行多元分析——定性资料的判别分析  

Reasonably carry out multivariate analysis:qualitative data discrimination analysis

作　　者：胡纯严 胡良平 Hu Chunyan;Hu Liangping(Graduate School,Academy of Military Sciences PLA China,Beijing 100850,China;Specialty Committee of Clinical Scientific Research Statistics of World Federation of Chinese Medicine Societies,Beijing 100029,China)

机构地区：[1]军事科学院研究生院,北京100850 [2]世界中医药学会联合会临床科研统计学专业委员会,北京100029

出　　处：《四川精神卫生》2024年第S1期19-23,共5页Sichuan Mental Health

摘　　要：本文目的是介绍与定性资料判别分析有关的基本概念、计算方法、两个实例以及SAS实现。基本概念包括定性变量与分类变量、判别分析、最大似然法、先验概率和先验分布、后验概率和后验分布;计算方法涉及最大似然判别法和Bayes公式判别法;两个实例的资料分别是“各类阑尾炎在各细目上出现的频率”以及“一例患者的症状和体征资料”;借助SAS,对两个实例的数据分别进行了判别分析,并对SAS输出结果做出了解释。The purpose of the paper was to introduce the basic concepts,calculation methods,two examples and SAS implementation related to the qualitative data discrimination analysis.Basic concepts included qualitative and categorical variables,discriminant analysis,maximum likelihood method,prior probability and prior distribution,posterior probability and posterior distribution.The calculation methods involved the maximum likelihood discriminant method and Bayes formula discriminant approach.The data in the two examples were“the frequency of various types of appendicitis in each detail”and“the symptoms and signs of a patient”.With the help of SAS software,the discriminant analysis was carried out on the data in the two examples,and an explanation was made for the output results of SAS.

关 键 词：定性资料 判别分析 最大似然法 先验概率 贝叶斯公式 

分 类 号：R195.1[医药卫生—卫生统计学]