Thread Subject: cftool options

I am trying to use the cftool options in MATLAB and I have
gotten confused by the syntax.

I am trying to change the options by writing the lines of
code that will change the default algorithm solver
from "Trust-Region" to "Gauss-Newton". I put in the code
to change that option but this is the error that I got.

s = fitoptions('Method','Gauss-Newton');
pl= cfit(f,162e-7,680e-9,10e-2,'options',s)

??? Error using ==> fitoptions>findmethod at 232
Method name "Gauss-Newton" is not valid.

Error in ==> fitoptions at 168
        token = findmethod(methodvalue);

Error in ==> PIT_no_wall at 66
s = fitoptions('Method','Gauss-Newton');


I am also trying to make 162e-7, 680e-9, and 10e-2 my
starting guess values for my three constants. Could
someone help my figure out what is wrong with my code.

--Ender--

"Ender " <jr147@msstate.edu> wrote in message
<g55o5g$7ru$1@fred.mathworks.com>...
> I am trying to use the cftool options in MATLAB and I
have
> gotten confused by the syntax.
>
> I am trying to change the options by writing the lines
of
> code that will change the default algorithm solver
> from "Trust-Region" to "Gauss-Newton". I put in the code
> to change that option but this is the error that I got.
>
> s = fitoptions('Method','Gauss-Newton');
> pl= cfit(f,162e-7,680e-9,10e-2,'options',s)
>
> ??? Error using ==> fitoptions>findmethod at 232
> Method name "Gauss-Newton" is not valid.
>
> Error in ==> fitoptions at 168
> token = findmethod(methodvalue);
>
> Error in ==> PIT_no_wall at 66
> s = fitoptions('Method','Gauss-Newton');
>
>
> I am also trying to make 162e-7, 680e-9, and 10e-2 my
> starting guess values for my three constants. Could
> someone help my figure out what is wrong with my code.
>
> --Ender--
>
>
 I figured out the answer to this problem. My next
question is how do I put each individual answer into its
own variable.

My code was:
Lemer = coeffvalues(c)

My output was:
Lemer =
    12 23 83


How do I take each number and put it into it's own
variable? Meaning:

a = 12
b = 23
c = 83


-Ender-

"Ender " <jr147@msstate.edu> wrote in message <g55qbf$4c6
$1@fred.mathworks.com>...
> "Ender " <jr147@msstate.edu> wrote in message
> <g55o5g$7ru$1@fred.mathworks.com>...
> > I am trying to use the cftool options in MATLAB and I
> have
> > gotten confused by the syntax.
> >
> > I am trying to change the options by writing the lines
> of
> > code that will change the default algorithm solver
> > from "Trust-Region" to "Gauss-Newton". I put in the
code
> > to change that option but this is the error that I got.
> >
> > s = fitoptions('Method','Gauss-Newton');
> > pl= cfit(f,162e-7,680e-9,10e-2,'options',s)
> >
> > ??? Error using ==> fitoptions>findmethod at 232
> > Method name "Gauss-Newton" is not valid.
> >
> > Error in ==> fitoptions at 168
> > token = findmethod(methodvalue);
> >
> > Error in ==> PIT_no_wall at 66
> > s = fitoptions('Method','Gauss-Newton');
> >
> >
> > I am also trying to make 162e-7, 680e-9, and 10e-2 my
> > starting guess values for my three constants. Could
> > someone help my figure out what is wrong with my code.
> >
> > --Ender--
> >
> >
> I figured out the answer to this problem. My next
> question is how do I put each individual answer into its
> own variable.
>
> My code was:
> Lemer = coeffvalues(c)
>
> My output was:
> Lemer =
> 12 23 83
>
>
> How do I take each number and put it into it's own
> variable? Meaning:
>
> a = 12
> b = 23
> c = 83
>
>
> -Ender-


I solved this problem as well Thanks though!

"Ender " <jr147@msstate.edu> wrote in message
<g55qrn$9hj$1@fred.mathworks.com>...
> "Ender " <jr147@msstate.edu> wrote in message <g55qbf$4c6
> $1@fred.mathworks.com>...
> > "Ender " <jr147@msstate.edu> wrote in message
> > <g55o5g$7ru$1@fred.mathworks.com>...
> > > I am trying to use the cftool options in MATLAB and
I
> > have
> > > gotten confused by the syntax.
> > >
> > > I am trying to change the options by writing the
lines
> > of
> > > code that will change the default algorithm solver
> > > from "Trust-Region" to "Gauss-Newton". I put in the
> code
> > > to change that option but this is the error that I
got.
> > >
> > > s = fitoptions('Method','Gauss-Newton');
> > > pl= cfit(f,162e-7,680e-9,10e-2,'options',s)
> > >
> > > ??? Error using ==> fitoptions>findmethod at 232
> > > Method name "Gauss-Newton" is not valid.
> > >
> > > Error in ==> fitoptions at 168
> > > token = findmethod(methodvalue);
> > >
> > > Error in ==> PIT_no_wall at 66
> > > s = fitoptions('Method','Gauss-Newton');
> > >
> > >
> > > I am also trying to make 162e-7, 680e-9, and 10e-2
my
> > > starting guess values for my three constants. Could
> > > someone help my figure out what is wrong with my
code.
> > >
> > > --Ender--
> > >
> > >
> > I figured out the answer to this problem. My next
> > question is how do I put each individual answer into
its
> > own variable.
> >
> > My code was:
> > Lemer = coeffvalues(c)
> >
> > My output was:
> > Lemer =
> > 12 23 83
> >
> >
> > How do I take each number and put it into it's own
> > variable? Meaning:
> >
> > a = 12
> > b = 23
> > c = 83
> >
> >
> > -Ender-
>
>
> I solved this problem as well Thanks though!


I have another problemm

I cannot figure out why the curvefitting tool will not use
my intial guess when trying to fit the curve that I
specified. I wrote the code so that I would input the same
initial conditions, but whenever I run the program. MATLAB
warns me that I have not specified intitial guesses so
MATLAB then chooses its own initial guess. Because of that
I get different coefficients whenever I run the program. I
need help writing the code to input the same starting
guess values.

Here is my code:

% Define the options for the curve fit & the independent
variable
s = fitoptions('Method','NonlinearLeastSquares',...
            'Algorithm','Gauss-Newton');
f = fittype('a+b*(1-exp(-
t/c))','independent','t','options',s);

% Displays the independent & dependent variable,
Coeffiecient names
indep = indepnames(f)
dep = dependnames(f)
coeffs = coeffnames(f)

% Establish first guess of Coefficients for algorithm
pl= fit(f,162e-7,680e-9,10e-2);

% Determine Coefficients
KO = fit(t,y,pl);
KO

I would like to state my initial guesses in the "pl"
variable, which I thought I was doing.

--Ender--

> I have another problemm
...
> pl= fit(f,162e-7,680e-9,10e-2);
>
> % Determine Coefficients
> KO = fit(t,y,pl);

Ender, in the first line above I'm not sure what you were trying to do. I
would think you would just want to assign your starting values to a vector:

pl= [162e-7,680e-9,10e-2];

In the last line, you have not included anything in the call to indicate
what kind of fit you want to make. I assume you intend to do something like
this:

fit(t,y,f,'start',p1)

-- Tom

"Tom Lane" <tlane@mathworks.com> wrote in message
<g57mco$8iv$1@fred.mathworks.com>...
> > I have another problemm
> ...
> > pl= fit(f,162e-7,680e-9,10e-2);
> >
> > % Determine Coefficients
> > KO = fit(t,y,pl);
>
> Ender, in the first line above I'm not sure what you
were trying to do. I
> would think you would just want to assign your starting
values to a vector:
>
> pl= [162e-7,680e-9,10e-2];
>
> In the last line, you have not included anything in the
call to indicate
> what kind of fit you want to make. I assume you intend
to do something like
> this:
>
> fit(t,y,f,'start',p1)
>
> -- Tom
>
>

What you suggested worked thanks!!

--Ender--