Least squares fit provides point estimates of parameters along with the covariance matrix over the parameters estimates. One can compute standard deviation on the individual parameters by taking a square root of the diagonal values of this matrix. One can use scipy’s curve_fit routine for this.

On the other hand, modeling the regression using a Bayesian approach outputs a distribution of parameters along with the mean, standard deviation and interval estimates (5%, 95%, 50% etc.).

Question - Can we compare the standard deviation of individual parameters from the curve_fit routine with the standard deviation obtained from the posterior distribution of parameters from Stan.