To the accuracy estimation of the high intensive flows of experimental data

The computational aspects of processing a large volume of experimental data related to the unsteadiness of the process, measurement inaccuracy, and inaccuracy of classifying algorithms are investigated. The limitations of the Bayesian approach to the problem of pattern recognition are also considered, when the maximum probability of matching the current state to one of the basic standards is determined by decomposing the fragment under study according to a known basis.