Re: Problem related to a linear regression
- From: MET <Marcel.E.Tschudin@xxxxxxxxx>
- Date: Fri, 02 Nov 2007 19:03:57 -0000
On 2 Nov., 18:20, Ray Koopman <koop...@xxxxxx> wrote:
In the first regression, you minimized error along the Y axis, but
in the second you switched to minimizing along the (rescaled) X axis.
Perhaps you should look into "orthogonal distance" regression,
which accomodates error on both axes simultaneously.
I think to see your point. The situation with the same physical
parameter on the x and y axis, only differing between observed (Y) and
estimated (Xf), makes me really confused. To me the y-values seem to
be 'fix'. Wouldn't it therefore be sufficient to find the x-values
(Xf) which best fit y=x? But then I'm again there what I have already
done by minimising (Y-Xf)^2 .... It really looks like I have to follow
your recommendation.
Thank you for your hint. Hopefully I manage to get this done with
Excel.
.
- References:
- Problem related to a linear regression
- From: MET
- Re: Problem related to a linear regression
- From: matt271829-news
- Re: Problem related to a linear regression
- From: MET
- Re: Problem related to a linear regression
- From: Ray Koopman
- Problem related to a linear regression
- Prev by Date: Financial Analyst Positions - NYC Corridor
- Next by Date: Re: GR9768 problems 23, 29, 32, 36, 37, 50, 63 (last set of questions before the test)
- Previous by thread: Re: Problem related to a linear regression
- Next by thread: The math symbol for Dirac delta function in multidimension.
- Index(es):
Relevant Pages
|