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]
st: new package smwoodbury
From
Sam Schulhofer-Wohl <[email protected]>
To
[email protected]
Subject
st: new package smwoodbury
Date
Tue, 3 Jan 2012 15:19:02 -0600
Thanks to Kit Baum, a new package -smwoodbury- is available for
download from SSC.
-smwoodbury- provides Mata functions that implement the
Sherman-Morrison-Woodbury formula for a rank-k update to a matrix
inverse and that use the formula to efficiently solve matrix
equations.
The formula is: inv(A+U*C*V) = inv(A) -
inv(A)*U*inv(inv(C)+V*inv(A)*U)*V*inv(A), where A is NxN, U is NxK, C
is KxK, and V is KxN.
Use of this formula can save computation time when N is much larger
than K and either inv(A) is easy to calculate or inv(A+U*C*V) must be
calculated many times for different values of U, C and V. Such
situations arise, for example, in Markov chain Monte Carlo estimation
of mixed models.
The package provides functions that are optimized for a variety of
special cases, such as U=V' or A=I. In addition, functions are
provided that efficiently solve the equation (A+U*C*V)*X = Y for X
without directly inverting A+U*C*V.
--
Sam Schulhofer-Wohl
Senior Economist
Research Department
Federal Reserve Bank of Minneapolis
90 Hennepin Ave.
Minneapolis MN 55480-0291
(612) 204-5484
[email protected]
Views expressed herein are those of the author and not necessarily
those of the Federal Reserve Bank of Minneapolis or the Federal
Reserve System.
*
* 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/