An increase in the accuracy of prediction of disasters by accounting for inhomogeneous data on a damage

摘要

A problem is considered for the union of statistical data on aftereffects of extreme situations. A likelihood criterion is suggested for recounting inhomogeneous data on a predictable year. The suggested technique permits reducing the statistical uncertainty of the prediction of the probability of disasters (catastrophes) that are found to be at the "tale" of the distribution of extreme situations with respect to a damage.