These are chat archives for astropy/astropy

Jul 2017
Thomas Spriggs
Jul 26 2017 17:29
For the LevMarLSQFitter, is the calculation to get "Param_cov" based on a variant of the chi-square test? As in finding a factor from the sum_sqrs / dof section, and then multiplying that with the current cov_x?
I ask as I am trying to first fit a 1D gaussian model, run the chi-square test to get some meaningful errors in Amplitude, then calc. Flux errors, handing this all off to a 2D moffat fitter and then running another chi-square test for a goodness of fit test.
But so far my values are either in the tens (40s or 50s etc.), in the range of 4-6, or the range of 0.0 to 0.1.