fminsearch combining maximum likelihood and least squares

Hello,

I got a question about fminsearch, I'm working on a curve fitting problem,and I would like to perform a minimization of -Loglikelihood and Least Squares.

Can I use fminsearch with a function like this:

[LogL, LS]=function(parameters)

I got two objective values.

Thanks for the help.

Ben

[Ben]

Help plz!

Answer 1

[Tom Lane]

If you use this function, fminsearch will try to minimize the first output. You could wrap "yourfunction" in a second function and that function could return the second output, or some combination of the two. Example (did not actually try to run it):

function out = wrapper(varargin)
[neglogl,ls] = yourfunction(varargin{:});
out = (neglogl+ls)/2;