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: PCA with unbalanced data
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: PCA with unbalanced data
Date
Wed, 6 Apr 2011 09:11:30 +0100
Well, perhaps it appears that PCA "kind of assumes" that, but using
this procedure certainly does. Need to check either way.
On Wed, Apr 6, 2011 at 3:01 AM, Stas Kolenikov <[email protected]> wrote:
> On Tue, Apr 5, 2011 at 3:05 PM, PINAR ERDEM <[email protected]> wrote:
>> I want to use PCA (principal componets analysis) with a dataset of 49 variables. However my data is unbalanced (unequal number of observations). My questions are if it is possible to run PCA with unbalanced data and how to get longest possible components/factors? Any suggestions would be very much appreciated.
>
> I think this is a perfect case for -mi impute mvn-. PCA kind of
> assumes normal data, so -mi- will come up with an appropriate
> estimator of the covariance matrix of the data; with some luck, you
> might even be able to pull it out of the guts of -mi- and analyze as
> is. If not, you can impute a few dozen times... and then get stuck, as
> -mi estimate- does not support -pca-.
>
*
* 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/