pint.fitter.fit_wls_svd

pint.fitter.fit_wls_svd(r: ndarray, sigma: ndarray, M: ndarray, params: List[str], threshold: float) → Tuple[ndarray, ndarray, ndarray, Tuple[ndarray, ndarray, ndarray]]

A utility function used by the WLS fitters.

Perform a linear WLS fit given timing residuals (r), uncertainties (sigma), and design matrix (M) using singular value decomposition.

To handle parameter degeneracies, singular values less than threshold are replaced with infinity, so that the fit only takes place in a non-singular subspace of the parameter space.

Returns the parameter deviations (dpars), parameter covariance matrix (Sigma), design matrix normalization factors (Adiag), and the results of the SVD.