[R-SIG-Finance] Portfolio optimization & PCA

[Gottlieb, Neil]

Jorge:

There is a lot of work on using PCA to reduce the correlation matrix.

It has been a very long time, since I did such work however...

Most do PCA and then to an promax rotation to make the resulting factor set
orthogonal. Also some using a sample technique such as EM to build the correlation
Matrix rather than do it all at once. 

Try a search on EM Applications. That firm has done such work. 
Also these 2 papers might be of interest.

This search string in google returns plenty: "promax rotation factor analysis"

Hope this helps a bit!

Neil

-----Original Message-----
From: r-sig-finance-bounces at stat.math.ethz.ch [mailto:r-sig-finance-bounces at stat.math.ethz.ch] On Behalf Of Jorge Nieves
Sent: Friday, December 04, 2009 10:17 AM
To: r-sig-finance at stat.math.ethz.ch
Subject: [R-SIG-Finance] Portfolio optimization & PCA

Hi,

I have a portfolio of securities that I like to determine the optimal
asset allocation using something simple as Markowitz (Mean-Variance). I
have the problem that some of the securities in the portfolio are highly
correlated. I would like to reduce my security set, or covariance matrix
before running the optimization, using something like PCA or clustering.
I was wondering if someone has done any work that might shed some like
on the process.? Is there piece of literature that some one can
recommend?

Thanks,

Jorge Nieves



	[[alternative HTML version deleted]]

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only.
-- If you want to post, subscribe first.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: StroynyRowe.pdf
Type: application/pdf
Size: 152000 bytes
Desc: StroynyRowe.pdf
URL: <https://stat.ethz.ch/pipermail/r-sig-finance/attachments/20091204/0cbfcbb3/attachment.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ACombinedLinearModel.pdf
Type: application/pdf
Size: 90849 bytes
Desc: ACombinedLinearModel.pdf
URL: <https://stat.ethz.ch/pipermail/r-sig-finance/attachments/20091204/0cbfcbb3/attachment-0001.pdf>