Mexico-Hat regression with Neural Net Fitting and Curve Fitting

Hi, i am rather new to matlab and still some clumsy.

I try to make regressions of known functions with different tools and toolboxes.

At the moment i try to get regression functions for the mexico hat function

 [X,Y] = meshgrid ( -10:.5:10 );
 R = sqrt (X.^2 + Y.^2 ) + eps;
 Z = sin(R) ./ R;

With the curve fitting Toolbox i get a good regression surface for the values. I just dont find where to get the underlying regression function? I want to use the regression function in further toolboxes.

With the NN Toolbox i made a regression for y=x^2. This works good, but now i have some problems how to make the input vector out of the mexico hut meshgrid. Am i right, that i need two input neurons, each for X and Y, Z is the target? Can someone give me a hint with the inputs?

Thanks

Frank

[Greg Heath]

I really do not understand what you are trying to do. Can you explain more?