Problem using nlinfit function

Hi everyone,

I am having trouble using nlinfit for making a regression with an own function.

I have the following script:

    xdata = 1:150; 
    xdata = xdata';
    ydata = beta;
    beta0 = [0 -1 20];
    n = 1:150;
    y = nlinfit(xdata,ydata,fun,beta0);.

and the function fun.m is this:

    function y = fun(beta0,n);
    y = beta0(1)+ beta(2).*exp(-n/beta(3));

The basic original function that I want to use is a +be^(t/tau) to fit a set of data points:

I have tried a lot, but I get different errors, variables not defined, input must be 4 variables etc. Please help me out!

Answer 1

[Torsten]

1. ydata has to be an array of the same size as xdata (I don't know the size of the array beta).

2. The call to nlinfit must read

y = nlinfit(xdata,ydata,@fun,beta0);

3. fun.m must read

 function y = fun(beta0,xdata);
 y = beta0(1)+ beta0(2)*exp(-xdata/beta0(3));

Best wishes

Torsten.

[Anthony Bendeler]

I had to use xdata instead of trial number array n, thank you! And I didnt't use @ for the function. Thanks a lot! It worked!

Answer 2

[Walter Roberson]

      y = beta0(1)+ beta0(2).*exp(-n/beta0(3));