Re: Curve fitting to known equation of line
- From: "John D'Errico" <woodchips@xxxxxxxxxxxxxxxx>
- Date: Fri, 24 Mar 2006 19:59:52 -0500
Duncan Blyth wrote:
I am trying to fit a best fit line to a data set, for which I know
the equation of the line. The coefficients (K,J,L) and omegad (a
fixed value of omega) in the equation are constant. I can generate
a
set of data, but I cannot work backwards using this equation to
produce a best fit. This is mainly due to the fact that there is an
inverse relation in the equation. The basis fitting does not
provide
any luck. To add to my problems, I am unable to find the value of V
experimentally around the area where omega==omegad. This could
partially ease my troubles.
I have a link to the equation here : <http://personalpages.manchester.ac.uk/student/d.blyth/equation2.bmp>
Do I need the curve fitting toolbox to aid me (I only have the
student version) or is there a method that I am overlooking, which
would allow me to find K,J&L?
You cannot (uniquely) find all three of K, J, & L.
Suppose there was some arbitrary best combination
of these three parameters? Note that 2*K, 4*J, & 4*L
would then be equally good.
So you can arbitrarily set any one of these three
parameters to have the value 1.
Having done so, then the optimization toolbox (or
even fminsearch) will help you to estimate the best
pair of the other two parameters.
HTH,
John D'Errico
.
- Follow-Ups:
- Re: Curve fitting to known equation of line
- From: Duncan Blyth
- Re: Curve fitting to known equation of line
- References:
- Curve fitting to known equation of line
- From: Duncan Blyth
- Curve fitting to known equation of line
- Prev by Date: Delta Symbol into Excel
- Next by Date: Re: Curve fitting to known equation of line
- Previous by thread: Curve fitting to known equation of line
- Next by thread: Re: Curve fitting to known equation of line
- Index(es):
Relevant Pages
|
