cfit extract normalization mean and std

Hello,

I am fitting a rational function to some data, in my case x varies over many orders of magnitude, so I used to 'Normalize' option when fitting. There is a normalization mean and std to the x, how do I extract that? There isn't a public method to cfit for this as far as I know.

I have attached the fit function in question. The values I am interest in are:

    where x is normalized by mean 6.21e+05 and std 1.65e+06

Answer 1

[Maeri]

In the cfit source code, the mean and std are private variables, and there exists no access function as for coeffvalues.

Thus I don't think there is a way to access mean and std. A work around is just taking mean and std of your xdata input.

Answer 2

[Tobias Held]

I would like to share my code for the answer from @Maeri, thank you I was looking for this!

Add this to your Fit function.

[fitresult,norm, gof] = createFitNom(xData, yData)
[...]
% Save Normalize values
[x,mean,std] = normalize(xData);
norm.mean = mean;
norm.std = std;
[...]
end

I am not yet sure if a struct is a good data type for this, but the idea should be clear now.