Article View: sci.math.num-analysis
Article #310Re: Optimization Question Two
From: Ajay Shah
Date: Fri, 17 Apr 1992 05:28
Date: Fri, 17 Apr 1992 05:28
22 lines
1315 bytes
1315 bytes
dow@nova2.stanford.edu (Keith Dow) writes:
>One last question. It seems to me that nonlinear optimization
>is nonlinear optimization is nonlinear optimization. So why
>are there different routines for fitting data to a nonlinear
>function and for optimizing a nonlinear function?
In the case of MLE, it turns out that we can exploit properties
of the objective function (the likelihood function) to gain
efficiency. Esp. once you have "decent" seeds, there are ways
to get big gains.
I believe there are similar tricks for NLS.
It's sortof like matrix inversion. Matrix inversion is matrix inversion
is matrix inversion. But the name of the game is exploiting
structure and specail properties.
-ans.
--
Ajay Shah, (213)749-8133, ajayshah@usc.edu
Technical Details
Message-ID:
<kusogmINN1ke@alhena.usc.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!usc!newsReferences:
<MCGRANT.92Apr16172518@isl.stanford.edu> <1992Apr17.033445.18482@morrow.stanford.edu> <1992Apr17.034224.18996@morrow.stanford.edu>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!usc!news From: ajayshah@alhena.usc.edu (Ajay Shah) Newsgroups: sci.math.num-analysis Subject: Re: Optimization Question Two Date: 16 Apr 1992 22:28:54 -0700 Organization: University of Southern California, Los Angeles, CA Lines: 22 Sender: ajayshah@alhena.usc.edu (Ajay Shah) Message-ID: <kusogmINN1ke@alhena.usc.edu> References: <MCGRANT.92Apr16172518@isl.stanford.edu> <1992Apr17.033445.18482@morrow.stanford.edu> <1992Apr17.034224.18996@morrow.stanford.edu> NNTP-Posting-Host: alhena.usc.edu