Re: Limitations of Curve Fitting Toolbox




Torsten Hennig <Torsten.Hennig@xxxxxxxxxxxxxx> wrote
in message
<2024760802.18599.1250866136633.JavaMail.root@gallium.
mathforum.org>...

Torsten Hennig <Torsten.Hennig@xxxxxxxxxxxxxx>
wrote
in message

<49178287.18261.1250862076099.JavaMail.root@xxxxxxxxxx
thforum.org>...
Hi,

In the technical documentation p 2-22 for
version
2
of the curve fitting toolbox it says that.

All selected variables must be of the same
length,
also you can only import vectors not matrices
and
scalars.

If there a way round this limitation?

I have one vector of independent variables
and
the
same number of dependent variables. However I
have a
series of the dependent varaibles and would
like
to
find the set of parameters that gives the
overall
best fit to these series of observed values.
Is
there
a way to do this using the curve fitting tool
box?

Just repeat the vector of independent variables
as
often as you have measured the dependent
variables
in your series.

Best wishes
Torsten.
Hi,

Would this give me one global best fit or a
series of
fits to each days data?

I really need one global best fit.

Thanks

Hugh

I suggested to use _all_ the data (with repeating
independent variables) in _one_ optimization.
This will of course give you a global best fit
(because you only get back one set of parameters).

Best wishes
Torsten.



Thank you so so much for this!!!

Can you please explain to me how to set the problem
up like this?

thanks

Hugh

Use lsqnonlin or lsqcurvefit from the optimization
toolbox.

Best wishes
Torsten.
.

Relevant Pages
Loading