Criteria of the validation of experimental and evaluated covariance dataS. Badikov
Central Research Institute of Management, Economics and Information (Atominform), 127434 Moscow, Russia
Published online: 21 May 2008
Criteria of the validation of experimental and evaluated covariance data are reviewed. In particular: a) criterion of the positive definiteness for covariance matrices, b) relationship between the "integral" experimental and estimated uncertainties, c) validity of the statistical invariants, d) restrictions imposed to correlations between experimental errors, are described. The examples of applying these criteria in practice are presented.
© CEA 2008