Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Total least squares with Mathematica?
381 views
Skip to first unread message
UMass74
May 15, 2012, 4:56:44 AM5/15/12
to
Hi Group:
I need to do some regression where both sets of data contains errors. I'm told total least squares handles this type of regression better.
I've RFD in Mathematica's site and done some Google searches and come up empty. Does anyone know of a source about using Mathematica to do this type of regression that I could look at?
Thanks for reading!
I need to do some regression where both sets of data contains errors. I'm told total least squares handles this type of regression better.
I've RFD in Mathematica's site and done some Google searches and come up empty. Does anyone know of a source about using Mathematica to do this type of regression that I could look at?
Thanks for reading!
Barrie Stokes
May 16, 2012, 4:28:57 AM5/16/12
to
Hi UMas74
I'm never ashamed to start with Wikipedia - very often the maths / stats articles are excellent, and there are always references.
In this case,
http://en.wikipedia.org/wiki/Errors-in-variables_models
HTH
Barrie
>>> On 15/05/2012 at 6:54 pm, in message <2012051508...@smc.vnet.net>,
I'm never ashamed to start with Wikipedia - very often the maths / stats articles are excellent, and there are always references.
In this case,
http://en.wikipedia.org/wiki/Errors-in-variables_models
HTH
Barrie
>>> On 15/05/2012 at 6:54 pm, in message <2012051508...@smc.vnet.net>,
Barrie Stokes
May 16, 2012, 4:29:27 AM5/16/12
to
Hi again
You might get some coding ideas from
http://vizsage.com/other/leastsquaresexcel/
If it can be done in Excel, ...
You might get some coding ideas from
http://vizsage.com/other/leastsquaresexcel/
If it can be done in Excel, ...
Darren Glosemeyer
May 16, 2012, 4:27:25 AM5/16/12
to
On 5/15/2012 3:54 AM, UMass74 wrote:
> Hi Group:
>
> I need to do some regression where both sets of data contains errors. I'm told total least squares handles this type of regression better.
>
> I've RFD in Mathematica's site and done some Google searches and come up empty. Does anyone know of a source about using Mathematica to do this type of regression that I could look at?
>
> Thanks for reading!
>
The orthogonal regression case in
> Hi Group:
>
> I need to do some regression where both sets of data contains errors. I'm told total least squares handles this type of regression better.
>
> I've RFD in Mathematica's site and done some Google searches and come up empty. Does anyone know of a source about using Mathematica to do this type of regression that I could look at?
>
> Thanks for reading!
>
http://demonstrations.wolfram.com/OrdinaryRegressionAndOrthogonalRegressionInThePlane/
solves the total least square problem. You could probably look at the
code for that demonstration as a starting point for you application.
Darren Glosemeyer
Wolfram Research
Ray Koopman
May 17, 2012, 4:10:00 AM5/17/12
to
that ran Feb 10-16, 2009
http://forums.wolfram.com/mathgroup/archive/2009/Feb/msg00258.html
https://groups.google.com/group/comp.soft-sys.math.mathematica/browse_frm/thread/a410b10426005c01
Hobbs, Sylvia (DPH)
May 17, 2012, 4:11:02 AM5/17/12
to
There is also an entire issue of the Journal of Computational Statistics and Data Analysis dedicated to total least squares and data error modeling. Here is an link to the opening editorial to that issue:
ftp://ftp.esat.kuleuven.be/pub/sista/ida/reports/07-119.pdf
Sylvia Hobbs
________________________________________
From: Barrie Stokes [Barrie...@newcastle.edu.au]
Sent: Wednesday, May 16, 2012 4:24 AM
Subject: Re: Total least squares with Mathematica?
Hi UMas74
I'm never ashamed to start with Wikipedia - very often the maths / stats articles are excellent, and there are always references.
In this case,
http://en.wikipedia.org/wiki/Errors-in-variables_models
HTH
UMass74
May 19, 2012, 5:41:59 AM5/19/12
to
Thanks to everyone for the suggestions! I'll check each reference.
0 new messages