There are many complementary strategies to estimate the extrapolation errors of a model calibrated in least-squares fits. We consider the Skyrme-Hartree-Fock model for nuclear structure and dynamics and exemplify the following five strategies: uncertainties from statistical analysis, covariances between observables, trends of residuals, variation of fit data, and dedicated variation of model parameters. This gives useful insight into the impact of the key fit data as they consist of binding energies, charge rms radii, and charge formfactor. Amongst others, we check in particular the predictive value for observables in the stable nucleus Pb-208, the super-heavy element (266)Hs, r-process nuclei, and neutron stars.
Type of Publication:
Journal article published