Re: Curve fitting of data from text file

[Przemek Klosowski]

On 03/24/2010 10:38 AM, socware wrote:

> I want to do curve fitting and get the value of the coefficients which gives
> me minimum error for the equation:
>
> I_d=a*(V_g-b)^2   %the value of I_d and V_g I get from the two columns of a
> text file. I have around 2000 values of I_d&  V_g, basically trying to fit
> the data from transistor I-V curve.
>
> I have a pseudo code which will give an idea what I want to do.
>
> i,j=0
> V_g= data from 2nd column of text file
> I_d= data from 1st column of text file
> for a=a0/10 to a0*10                # the value of a0 can be between some
> range like 100 to 10000
>    i=i+1
>    for b=b0/10 to b0*10              # the value of b0 can be between some
> range like 30 to 3000
>      j=j+1
>      I_dfit(1:size(V_g))=a*(V_g-b)^2     # functione to get best fit? if
> polyfit then how do I implement here.
>      error(i,j)=sum( (I_fit(1:size(V_g))-I_d(1:size(V_g)))^2 )
>    end
> end
> bestIandJ=findminimum(error)
>
>
>
> Any suggestion how do I proceed.?
>
> Thanks

leasqr(V_g,I_d,[1 1],@(x,p)(p(1).*(x-p(2)).^2))

might give you an approximation; [1 1] is the
initial guess for a and b.