APPLICATION OF THE BAYESIAN-ESTIMATION METHOD TO THE EXAFS DATA-ANALYSIS

The Bayesian estimation method which is based on mathematical statistics combines experimental signals and pre-information in a common loss function. We have improved the original version of the Bayesian method and can take into account implicit constraints between the parameters of the EXAFS function, that must be fulfilled only within chosen error bounds. These linear relations connect occupation probabilities or phase fractions, coordination numbers, interatomic distances and mean-square relative displacements of the atoms.