How do I define an output vector as real and positive

Hi all

I am currently running this script to fit a complex function on my data using lsqcurvefit.

clear all;

hold off E1=importdata('E1.mat'); %E1 and DR1 are the names of my arrays

DR=importdata('DR1.mat');

hold on plot(E1,DR)

%x(1)=C

%x(2)=f

%x(3)=Eg

%x(4)=G

x0= [1 0 1 0.02 ];

F=@(x,E1)x(1)*exp(1i*x(2))./((E1-x(3)-1i*x(4))^2);

[x,resnorm,~,exitflag,output] = lsqcurvefit(F,x0,E1,DR);

plot(E1,F(x,E1),'--r')

hold off

It is working perfectly except for the part where all the fitted parameters are returned as complex numbers when they should be real and positive. Is there some way to fix this?

Thanks in advance

[Star Strider]

I don’t know what you are doing, but (from the Euler identities), exp(1i*x(2)) will be evaluated essentially as cos(x(2)) + i*sin(x(2)), and you are also multiplying x(4)^2 by 1i. If you do not want complex parameters, you will likely have to re-write your model. I would go to the literature to see what others have done in similar situations.

Answer 1

[Roger Wohlwend]

The same question occured recently on Matlab answers: Link to the question.