Prev: There may be a singularity in the solution
Next: mc-cdma
From: ignacio vilchis on 8 Feb 2005 00:45
Hi, I am trying to fit a vonbertalanffy growth curve to some fish
growth data, using iterative least squares and the fminsearch
function.
The VonB function has three unknowns hence you need three initial
guesses. I can assign the function to a handle ok with this command:
VonB = @(Lmax,b,k) sum(((Lmax*(1-b*exp(-k*age)))-(real data)).^2)
and this for the initial guess for the three parameters
x0 = [100 .8 .5]
but when I call for the fminsearch to find the minimun values
x = fminsearch(VonB,x0)
I get this, when I did define b? Any ideas?

??? Input argument "b" is undefined.
Error in ==> fminsearch at 175
fv(:,1) = funfcn(x,varargin{:});

Thanks in advance for any help.
From: Peter Spellucci on 8 Feb 2005 07:43

In article <eefb168.-1(a)webx.raydaftYaTP>,
"ignacio vilchis" <lvilchis(a)ucsd.edu> writes:
>Hi, I am trying to fit a vonbertalanffy growth curve to some fish
>growth data, using iterative least squares and the fminsearch
>function.
>The VonB function has three unknowns hence you need three initial
>guesses. I can assign the function to a handle ok with this command:
>VonB = @(Lmax,b,k) sum(((Lmax*(1-b*exp(-k*age)))-(real data)).^2)
>and this for the initial guess for the three parameters
>x0 = [100 .8 .5]
>but when I call for the fminsearch to find the minimun values
>x = fminsearch(VonB,x0)
>I get this, when I did define b? Any ideas?
>
>??? Input argument "b" is undefined.
>Error in ==> fminsearch at 175
>fv(:,1) = funfcn(x,varargin{:});
>
>Thanks in advance for any help.

the parameter of VonB must be a vector with three components, not three scalars
since fminsearch takes one vector as optimization variable and treats
other variables in a function called as "additional parameters".
hth
peter
From: Ignacio Vilchis on 8 Feb 2005 14:40
I thought that making the parameters as
x0 = [100 .8 .5]
made that into a vector having three components, that is what I have,
but still get the same error message. Are you reffering to something
else?

Peter Spellucci wrote:
>
>
> In article <eefb168.-1(a)webx.raydaftYaTP>,
> "ignacio vilchis" <lvilchis(a)ucsd.edu> writes:
> >Hi, I am trying to fit a vonbertalanffy growth curve to some
fish
> >growth data, using iterative least squares and the fminsearch
> >function.
> >The VonB function has three unknowns hence you need three
initial
> >guesses. I can assign the function to a handle ok with this
> command:
> >VonB = @(Lmax,b,k) sum(((Lmax*(1-b*exp(-k*age)))-(real
data)).^2)
> >and this for the initial guess for the three parameters
> >x0 = [100 .8 .5]
> >but when I call for the fminsearch to find the minimun values
> >x = fminsearch(VonB,x0)
> >I get this, when I did define b? Any ideas?
> >
> >??? Input argument "b" is undefined.
> >Error in ==> fminsearch at 175
> >fv(:,1) = funfcn(x,varargin{:});
> >
> >Thanks in advance for any help.
>
> the parameter of VonB must be a vector with three components, not
> three scalars
> since fminsearch takes one vector as optimization variable and
> treats
> other variables in a function called as "additional parameters".
> hth
> peter
>
From: Marcelo Marazzi on 8 Feb 2005 17:44
Fminsearch expects the objective to depend on a vector
variable instead of on separate scalar variables. (I call
this vector variable "x" below, but one can use any other,
more descriptive, variable name.)

VonB = @(x) sum(((x(1)*(1-x(2)*exp(-x(3)*age)))-(real_data)).^2)
x0 = [100 .8 .5]
x = fminsearch(VonB,x0)

-marcelo

Ignacio Vilchis wrote:
> I thought that making the parameters as
> x0 = [100 .8 .5]
> made that into a vector having three components, that is what I have,
> but still get the same error message. Are you reffering to something
> else?
>
> Peter Spellucci wrote:
>
>>
>>In article <eefb168.-1(a)webx.raydaftYaTP>,
>>"ignacio vilchis" <lvilchis(a)ucsd.edu> writes:
>>
>>>Hi, I am trying to fit a vonbertalanffy growth curve to some
>
> fish
>
>>>growth data, using iterative least squares and the fminsearch
>>>function.
>>>The VonB function has three unknowns hence you need three
>
> initial
>
>>>guesses. I can assign the function to a handle ok with this
>>
>>command:
>>
>>>VonB = @(Lmax,b,k) sum(((Lmax*(1-b*exp(-k*age)))-(real
>
> data)).^2)
>
>>>and this for the initial guess for the three parameters
>>>x0 = [100 .8 .5]
>>>but when I call for the fminsearch to find the minimun values
>>>x = fminsearch(VonB,x0)
>>>I get this, when I did define b? Any ideas?
>>>
>>>??? Input argument "b" is undefined.
>>>Error in ==> fminsearch at 175
>>>fv(:,1) = funfcn(x,varargin{:});
>>>
>>>Thanks in advance for any help.
>>
>>the parameter of VonB must be a vector with three components, not
>>three scalars
>>since fminsearch takes one vector as optimization variable and
>>treats
>>other variables in a function called as "additional parameters".
>>hth
>>peter
>>
From: ignacio vilchis on 8 Feb 2005 17:53
thanks marcelo
got it!!

Marcelo Marazzi wrote:
>
>
> Fminsearch expects the objective to depend on a vector
> variable instead of on separate scalar variables. (I call
> this vector variable "x" below, but one can use any other,
> more descriptive, variable name.)
>
> VonB = @(x) sum(((x(1)*(1-x(2)*exp(-x(3)*age)))-(real_data)).^2)
> x0 = [100 .8 .5]
> x = fminsearch(VonB,x0)
>
> -marcelo
>
> Ignacio Vilchis wrote:
>> I thought that making the parameters as
>> x0 = [100 .8 .5]
>> made that into a vector having three components, that is what I
> have,
>> but still get the same error message. Are you reffering to
> something
>> else?
>>
>> Peter Spellucci wrote:
>>
>>>
>>>In article <eefb168.-1(a)webx.raydaftYaTP>,
>>>"ignacio vilchis" <lvilchis(a)ucsd.edu> writes:
>>>
>>>>Hi, I am trying to fit a vonbertalanffy growth curve to
some
>>
>> fish
>>
>>>>growth data, using iterative least squares and the
fminsearch
>>>>function.
>>>>The VonB function has three unknowns hence you need
three
>>
>> initial
>>
>>>>guesses. I can assign the function to a handle ok with
this
>>>
>>>command:
>>>
>>>>VonB = @(Lmax,b,k) sum(((Lmax*(1-b*exp(-k*age)))-(real
>>
>> data)).^2)
>>
>>>>and this for the initial guess for the three parameters
>>>>x0 = [100 .8 .5]
>>>>but when I call for the fminsearch to find the minimun
values
>>>>x = fminsearch(VonB,x0)
>>>>I get this, when I did define b? Any ideas?
>>>>
>>>>??? Input argument "b" is undefined.
>>>>Error in ==> fminsearch at 175
>>>>fv(:,1) = funfcn(x,varargin{:});
>>>>
>>>>Thanks in advance for any help.
>>>
>>>the parameter of VonB must be a vector with three
components, not
>>>three scalars
>>>since fminsearch takes one vector as optimization variable
and
>>>treats
>>>other variables in a function called as "additional
parameters".
>>>hth
>>>peter
>>>
>
 | 
Pages: 1
Prev: There may be a singularity in the solution
Next: mc-cdma