#1 主成份分析和因子分析简介
主成份分析和因子分析用于将数据中多个相关的变量合并为少数几个潜在的维度(underlying dimensions)。Stata中相关命令主要包括:
- pca: principle components analysis,主成分分析
- factor:因子分析,用于提取不同类型的因子
- screeplot:根据pca或factor画出碎石图(scree graph,也叫特征值标绘图)
- rotate:使用factor命令之后,进行正交或斜交旋转
- predict:在使用pca、factor和rotate命令之后,创建因子分或符合变量。便于下一步进行建模分析
- alpha:哥伦巴哈阿尔法信度系数。如果不使用因子分析和主成份分析,而是直接将相关变量相加,则需要检验它们的alpha系数
- cluster:聚类分析
#2 例子:planets数据分析
第一步:导入数据
. cd "D:StataStatistics with STATA"
. use "D:StataStatistics with STATAplanets.dta", clear
第二步:查看数据
. des
/*
Contains data from D:StataStatistics with STATAplanets.dta
obs: 9 Solar system data
vars: 12 2 Jul 2012 06:11
---------------------------------------------------------------------------------------------
storage display value
variable name type format label variable label
---------------------------------------------------------------------------------------------
planet str7 %9s Planet
dsun float %9.0g Mean dist. sun, km*10^6
radius float %9.0g Equatorial radius in km
rings byte %8.0g ringlbl Has rings?
moons byte %8.0g Number of known moons
mass float %9.0g Mass in kilograms
density float %9.0g Mean density, g/cm^3
logdsun float %9.0g natural log dsun
lograd float %9.0g natural log radius
logmoons float %9.0g natural log (moons + 1)
logmass float %9.0g natural log mass
logdense float %9.0g natural log dense
---------------------------------------------------------------------------------------------
Sorted by: dsun
Note: Dataset has changed since last saved.
*/