How can i estimate the uncertainty for the parameters returned by fit function?

Hello everyone!

I have to calculate the Stokes parameters from a fit of the experimental data. To do this i have, like first thing, implemented the fit of the data with the function "fit" and a custom model fitting function. Now, the program returns to me the parameters, but i have to calculate the Stokes ones from these with relative errors. To do this, i need the uncertainties of the fit parameters to propagate the erros on the Stokes' parameters: how can i estimate these?

Thousands thanks for the patient and the answers

Answer 1

[Matt J]

You could use confint to extract confidence bounds on the parameters, e.g.,

https://www.mathworks.com/help/curvefit/cfit.confint.html#f3-529841

[Sterling Baird]

Thoughts for higher-dimensional fits?

[Matt J]

Is there an issue with higher dimensional fits?