Today we looked at how to do non-linear regression. The key is to (1) define an objective function of fitting parameters which is a minimum when you have the best fit values (usually the norm of the residual), and (2) scale the model parameters so that they are all of roughly the same magnitude. This conditioning can be critical as otherwise function minimization routines have a hard time converging to the correct value! As an example, we look at the senior lab catalytic oxidation experiment given here. You will also need to download the objective function delta.m.