From: Michael Stevens (Michael.Stevens_at_[hidden])
Date: 2003-07-24 14:34:26
A quick last minute contribution to this thread.
There was a fair bit of incidental information in the discussion. If
anyone wants to formalise things a little further one good place to put
is the Boost Wiki. There is already a page there regarding uBLAS and a
page on uBLAS and Linear Algebra which could be usefuly updated and
extended.
David originally asked about the following techniques.
LU factorization
Cholesky factorization
Gaussian Elimination
Ordinary Least Squares Regression
Most of these are well covered by the current set of uBLAS binding to
Atlas. Going a little further from Least Squares to general estimation
techniques it may be worth looking at my one Bayes++ library. This is
uBLAS based and does all the linear algebra work for a wide range of
estimation techniques. Least squares can be simply implemented as
stationary model. See http://bayesclasses.sourceforge.net/
Michael