Message-ID: <31967454.1075840778372.JavaMail.evans@thyme>
Date: Thu, 12 Apr 2001 11:38:00 -0700 (PDT)
From: zimin.lu@enron.com
To: zhiyong.wei@enron.com
Subject: correlation matrix reduction
Cc: stinson.gibner@enron.com, tanya.tamarchenko@enron.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Bcc: stinson.gibner@enron.com, tanya.tamarchenko@enron.com
X-From: Zimin Lu <Zimin Lu/HOU/ECT@ECT>
X-To: Zhiyong Wei <Zhiyong Wei/HOU/ECT@ECT>
X-cc: Stinson Gibner <Stinson Gibner/HOU/ECT@ECT>, Tanya Tamarchenko <Tanya Tamarchenko/HOU/ECT@ECT>
X-bcc: 
X-Folder: \vkamins\c:\technote\mail\Techmemos
X-Origin: KAMINSKI-V
X-FileName: vincent kaminski 1-30-02.pst



Zhiyang,

Got your message.  

As we discussed, the best way to breakdown the big correlation matrix between 
different location indices is through "cluster analysis".  The idea is to select major
hubs and the "satellite" markets.  By establish the correlation between the hubs
and correlation between the satellites and the hubs, the big matrix can be significantly
reduced.    Then the traders only need to input the correlations between the hubs.

This technique has been applied in our Value at Risk system.  You may talk to 
Tanya to find out the details. 


Zimin




PS: the wsprd code you requested.

 