nonlinear least squares fit

Hi everyone,

sorry, but I am trying to fit some data and don't get where I am going wrong. I have defined my function

function F = nonlinfun(x,t)
F = x(1)*t.^2 + x(2)*t + x(3) + x(4)*sin(2*pi*(t+x(5)));

and then I am trying to do the fit

clear;

read = importdata('c:/data.txt',',');
data = read(:,7);

n = length(data);
t = 1:n;
x0 = rand(5,1);
x0 = 300;

[x,resnorm] = lsqcurvefit(@nonlinfun,x0,t,data);

The data I am trying to fit can be found here:

http://scrippsco2.ucsd.edu/data/in_situ_co2/monthly_mlo.csv

(coumn 7) if anyone is interested. I saved it in a file 'data.txt' and reading that out works. When I try to run above sequence Matlab returns the following error:

Function value and YDATA sizes are incommensurate.

I got the simple Matlab examples to work but have tried unsuccessfully all day with the atmospheric data. I guess there is something simple I am overlooking?

Thanks for the help.

Answer 1

[Ivan van der Kroon]

t is a row-vector while data is a column-vector. Change it to

t=(1:n)';

Btw, I got

x=  0.0012
  0.1046
 -7.6955
 -2.0487
 -3.2998

Answer 2

[Sebastian Schnelle]

Thanks. I get different results but that might have to do with slightly different starting values. Thanks again.