Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: creating a foreach loop for principal components


From   Nick Cox <[email protected]>
To   [email protected]
Subject   Re: st: creating a foreach loop for principal components
Date   Fri, 2 Sep 2011 19:20:42 +0100

Your text talks initially about omitting observations, but your code
examples are entirely in terms of omitting variables. I am going to
assume you mean the latter.

NB: Observations are rows; variables are columns. BTW, -pca- is a
command, not a function.

The structure you seem to be asking for can be coded like this.

unab all : var1-var300
tokenize `all'

forval j = 1/300 {
      local this : list all - ``j''
      pca `this'
      predict pc1-pc5
      regress ``j'' pc1-pc5
     drop pc1-pc5
}

Nick


On Fri, Sep 2, 2011 at 7:09 PM, Dmitriy Glumov <[email protected]> wrote:

> I am working on principal components and have run into a problem
> creating a loop. I have about 300 observations and each one needs to
> undergo a 3 step process: 1) Creating principal components WITHOUT the
> respective observation (principal components function itself is very
> simple, but I struggle to create a loop for it), 2) Predicting first 5
> components (this step doesn't change, but component values themselves
> are dependent on step 1), and 3) Regressing the respective variable
> onto the components.
>
> To illustrate, the process for variable 1 would look something like this:
> pca var2-var300
> predict pc1-pc5
> regress var1 pc1-pc5
>
> For variable 2, it would be this:
>
> pca var1 var3-var300
> predict pc1-pc5
> regress var2 pc1-pc5
>
> For variable 3,
>
> pca var1-var2 var4-var300
> predict pc1-pc5
> regress var3 pc1-pc5
>
> and, so on for the remaining variables. It is certainly doable by
> hand, but the process is very time-consuming. I am sure there is a way
> to create a foreach loop but I just can't figure out how. So if you
> can help me with this, it would be greatly appreciated. Thank you for
> your consideration.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index