[Date Index] [Thread Index] [Author Index]

Re: Total least squares with Mathematica?

• To: mathgroup at smc.vnet.net
• Subject: [mg126514] Re: Total least squares with Mathematica?
• From: Darren Glosemeyer <darreng at wolfram.com>
• Date: Wed, 16 May 2012 04:23:17 -0400 (EDT)
• Delivered-to: l-mathgroup@mail-archive0.wolfram.com
• References: <201205150854.EAA10996@smc.vnet.net>

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

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

• References:
  • Total least squares with Mathematica?
    • From: UMass74 <fhasmith@frankandluann.net>