| Re: Robust curve fitting (BUG) [message #12439] |
Wed, 05 August 1998 00:00 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
By request, the fitting routines are now available in single packaged
archives. Both GZIPed tar files and ZIP archives are available. The
web page address is the same.
Another note: when I fixed the bug mentioned in a previous post, I
also added the PERROR keyword, which returns parameter errors
explicitly. (Previously you had to compute errors yourself using the
covariance matrix).
On the whole, the MINPACK-1 routines themselves are well-tested.
However, the original MINPACK-1 routines did not permit parameter
freezing or simple bounding constraints, nor did it automatically
calculate the parameter errors. These are my own additions; therefore
there may be a small amount of bug-ironing to get it just right.
Suggestions are welcome.
Craig
Craig Markwardt <craigmnet@astrog.physics.wisc.edu> writes:
>>>>
>>>> I recently had an opportunity to translate the MINPACK-1 curve-fitting
>>>> package into IDL. MINPACK is a minimization package available from
>>>> netlib, and has an excellent reputation. I have found that it is much
>>>> more robust, able to cope with singular matrices, etc. Since people
>>>> have been requesting, I polished it up a little bit, and am making it
>>>> available via my IDL web page:
>>>>
>>>> http://astrog.physics.wisc.edu/~craigm/idl/idl.html
>>
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@astrog.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: Robust curve fitting (BUG) [message #12450 is a reply to message #12439] |
Tue, 04 August 1998 00:00  |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
I found a small bug in MPFIT.PRO. It only bites in certain
circumstances. Specifically, if you fix some parameters, and place
bounding constraints on other parameters, then the fit procedure can
finish too soon, before the parameters have converged.
A new corrected version is on my web page, listed below.
Craig
>>>
>>> I recently had an opportunity to translate the MINPACK-1 curve-fitting
>>> package into IDL. MINPACK is a minimization package available from
>>> netlib, and has an excellent reputation. I have found that it is much
>>> more robust, able to cope with singular matrices, etc. Since people
>>> have been requesting, I polished it up a little bit, and am making it
>>> available via my IDL web page:
>>>
>>> http://astrog.physics.wisc.edu/~craigm/idl/idl.html
>
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@astrog.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
