Path: news.mathworks.com!not-for-mail
From: "Alexander" <alexander.erlich@gmail.com>
Newsgroups: comp.soft-sys.matlab
Subject: Fitting in Matlab
Date: Fri, 27 Mar 2009 11:03:01 +0000 (UTC)
Organization: The MathWorks, Inc.
Lines: 11
Message-ID: <gqibp5$5dm$1@fred.mathworks.com>
Reply-To: "Alexander" <alexander.erlich@gmail.com>
NNTP-Posting-Host: webapp-05-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1238151781 5558 172.30.248.35 (27 Mar 2009 11:03:01 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Fri, 27 Mar 2009 11:03:01 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 1217178
Xref: news.mathworks.com comp.soft-sys.matlab:528147


Hello, 

I use a student version of Matlab R2007a, and I have a question about fitting. The student version comes without the Curve Fitting Toolbox. However, there is a way of intuitive GUI fitting, but it's very limited: In my version, if I go to Tools --> Basic fitting (in any figure window), I can for example choose linear fitting. Unfortunately, only one dataset can be fitted simultaneously, and there is no way to exclude bad points from the fit.

I have to do a lot of fitting regularly (mostly linear regression), and in most problems I require to fit several datasets in one figure, as well as to be able to exclude points in a dataset from fitting. Do you know a good way to do that (apart from coding every fit by hand)? 

Matlab is really wonderful for manipulating data (recorded in physical experiments), it's also wonderful for plotting, so I guess there must be some way do some good fitting with it. I'd really hate to have to change to other tools like Origin oder Igor Pro, which by the way do not even have native Linux versions. 

That's why I'm really curious about your replies!

Alexander