Background estimation using a robust Bayesian analysis

A novel method for the estimation of the background in a powder diffraction pattern has been developed using a robust Bayesian analysis. In formulatin g a probabilistic approach to background fitting, the diffraction peaks are considered to be nuisance data that must be taken into account. The underl ying probability theory is discussed in terms of going beyond the Gaussian approximation normally associated with counting statistics and least-square s analysis. Various examples are presented that illustrate the general appl icability of this approach.