Question about Goodness of Fit

Hello,

I've been using nlinfit using a Boltzmann equaion to fit a simoid to my data. I'm trying to determine the goodness of fit and it seems there are a varity of ways to do this and I'm not exactly sure what to do. It seems R^2 is not longer accepted in terms of non-linear fits. Can anyone help out?

Also, I'm kind of confused as to what to do with the extra nlifit output, other than beta.

Can someone explain in layman's terms what each variable means? [beta,R,J,CovB,MSE,ErrorModelInfo]

The image is the fit to my data - the larger circles are the mean of the smaller circles at each x value.

Answer 1

[Athul Prakash]

Please refer to the documentation on nlinfit which explains each of the output arguments in depth.

https://in.mathworks.com/help/stats/nlinfit.html

Furthermore, the Statistical Toolbox provides many ways to asses Goodness of Fit, all of which is once again outlined in the documentation:

https://in.mathworks.com/help/curvefit/evaluating-goodness-of-fit.html