Fitting data using lsqnonlin

I want to fit a vector (1*4000) to a normal-gamma-gamma mixture model (summation of these three distribution) using lsqnonlin,'Algorithm:trust-region-reflective' (I have to use this command according to my reference). However, with different initial point, I got the message: "Local minimum found". I've read the MatLab suggestions about how this problem can be solved, but I can't figure it out. Your help would be greatly appreciated. Thanks

[Alan Weiss]

I am sorry, but I do not understand your question. Is the issue that you got a local minimum rather than a global minimum? If so, you should know that lsqnonlin only reports that it finds local solutions, it has no way of knowing whether or not it found a global solution. See Local vs. Global Optima.

If your question is about something else, please ask again, stating more explicitly what the issue is.

Alan Weiss

MATLAB mathematical toolbox documentation