When assigning a PDF (probability density function) to a parameter encountered in risk assessment models, one is often faced with the problem of uncertain estimates of the parameters of the PDF itself due to lack of measurements. Often there exist values from literature or measurements of analogues, but the exclusive use of these values ignores information in the measurements.

BABAR uses Bayesian statistics to derive distributions of parameters by taking measurements as well as generic values or measurements of analogues into account. The generic values or measurements of analogues are used to derive a prior distribution which, by application of Bayes' theorem, is updated with the new measurements. The resulting distribution, called the posterior distribution, is a compromise of the prior distribution and the measurements.