Curve fitting a complex function

I am trying to fit this function on my experimental data and receive the values of G and f. My data ought to look like this.

DR=exp(1i*f)/(E-Eg-1i*G)

Where Eg=1.18, DR and E are my measurements as arrays. f is the phase, E, Eg and G are all measured in eVs.

I am trying to utilize the LSQNONLIN function.

My question is: How do I have to define the function that needs to be inserted in LSQNONLIN in order to make it work?

[Christos]

Thank you,

I currently have these 2 scripts

1)fn.m

function deltaR1 = fn(E,G)

deltaR1=1/(E-1.18-1i*G);

and 2) fitting.m

E=E1;%E1 and DR1 are the names of my arrays

DR=DR1;

ft=lsqcurvefit('fn(E,G)');

f=fit(E,DR,ft);

A=coeffvalues(f);

fplot(f)

Matlab returns this error:

Error using lsqcurvefit (line 167) The input to LSQCURVEFIT should be either a structure with valid fields or consist of at least four arguments.

Error in fitting (line 6) ft=lsqcurvefit('fn,E,G');

[Matt J]

The way you are invoking lsqcurvefit does not resemble any syntax in the documentation...

http://www.mathworks.com/help/optim/ug/lsqcurvefit.html

Answer 1

[Matt J]

I think you should use lsqcurvefit instead. Other than that, see this recent related thread for dealing with complex valued functions,

http://www.mathworks.com/matlabcentral/answers/155262-how-to-specify-limits-for-lsqnonlin