This program carries out curve fitting of two variables X and Y, whose interdependence is described by a known model function, Y=f(X).
Parameters of this fitting function are estimated through maximising the adjoined likelihood function. The procedure reminds of weighted least square fitting but comprises some additional features:
1. It allows for statistically correlated variables - cov(X,Y) non-zero.
2. If the model function is non-linear in the variable X conventional least square fitting results in biased estimates of the parameters if there is a statistical correlation of X and Y. This is not the case for the fitting
program presented here.
The uncertainty of the estimated parameters may also be calculated. For this the method of Monte Carlo simulations is used. Requirements:
· MATLAB Release: R10
· Optimization Toolbox,Statistics Toolbox