Hi,

you might want to try EstimatedDistribution instead. I don't have your data, but in the documentation they do fit a multivariate distribution.

bndata = RandomVariate[BinormalDistribution[{1, 2}, {1/3, 4}, 3/4], 1000];
dist = EstimatedDistribution[bndata, BinormalDistribution[{Subscript[\[Mu], 1], Subscript[\[Mu], 2]}, {Subscript[\[Sigma], 1], Subscript[\[Sigma], 2]}, \[Rho]]]

Cheers,

Marco

I suspect that the initial issue with NonlinearModelFit is that your data is not in the correct format. The following might be considered:

data = Import["Z2.dat"];
z2 = Flatten[Table[{x, y, data[[x, y]]}, {x, 202}, {y, 101}], 1];

This gets you a 20,402 x 3 array (which is what NonlinearModelFit expects: {x,y,z} are the 20,402 "data points") rather than a 202 x 101 array. I would also suggest keeping your variables in lowercase (as all of Mathematical functions start in uppercase) and avoiding subscripts (unless there is some very special need for formatting purposes).

I've just used the indices for the values of x and y as I can't tell what those should be from your dataset.

Yes. But I would also get the sum of my data "z2c" to compare it with the integration of the fit, to know the difference.

Thnx for the answers. Those are pixel values and here is 202x202 data: http://expirebox.com/download/4e5038c0edae938ffdcdde24e10ec191.html

Thanks, here's my data: http://expirebox.com/download/001c9c77bde9bf47238279b3c81694bd.html Somehow your suggestion didn't work.