| Generalized Least Squares? [message #60353] |
Sun, 18 May 2008 19:26 |
Gernot Hassenpflug
Messages: 18
Registered: April 2001
|
Junior Member |
|
|
Dear all,
I'm involved in ongoing research on a problem that I solved to a first
approximation with weighted least squares (using MPFIT) but for a real
solution I require generalized least squares: WLS uses a diagonal
covariance matrix, i.e., the data errors are uncorrelated; GLS uses a
full covariance matrix, i.e., the data errors can be correlated.
I have not found any ready solution in IDL yet, and I am under the
impression that there is no analytical solution to GLS, so fairly
complicated numerical methods are required.
I have actually found a routine in MATLAB called "lscov" which can
solve this problem, and wonder whether there is a chance to hobble
something together in IDL? I'd be happy to try and modify MPFIT to be
able to deal with GLS too, could anyone comment on possibilities?
Best regards,
Gernot Hassenpflug
--
BOFH excuse #51:
Cosmic ray particles crashed through the hard disk platter
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
