我在做主成分分析时候,也碰到类似问题
找遍了论坛 ,貌似发现这两个帖子
https://bbs.pinggu.org/thread-2525560-3-1.html
https://bbs.pinggu.org/thread-2441468-1-1.html
主要观点是
选择所有特征根(Eigenvalue)>1 的成分(Comp),如果有3个,那么看那3个成分所对应的:Proportion:proportion1,proportion2,proportion3.
再用代码:predict f1 f2 f3
最后得出综合指标的代码:gen index=proportion1*f1+proportion2*f2+proportion3*f3
index就是你要的综合指标了。
但是我认为这种方法有问题
因为最后得到的f1 f2 f3本身是完全不相关的
加权求和 没有意义啊 我查了国外做法
是可以直接用第一个的,比如
https://stats.stackexchange.com/questions/133492/creating-a-single-index-from-several-principal-components-or-factors-retained-fr
https://www.statalist.org/forums/forum/general-stata-discussion/general/1393259-using-pca-to-make-an-index-for-a-new-variable
Under the terms of the game, the first PC is by definition the best one to serve as a summary.
[backcolor=rgba(252, 251, 248, 0.901961)]PCs are uncorrelated by definition. Therefore, as variables, they don't duplicate each other's information in any way. That means that there is no reason to create a single value (composite variable) out of them. Or, sometimes multiplying them could become of interest, perhaps - but not summing or averaging.[backcolor=rgba(252, 251, 248, 0.901961)]
[backcolor=rgba(252, 251, 248, 0.901961)]所以我建议就用第一个



雷达卡



京公网安备 11010802022788号







