Оценки вероятности ошибки в байесовской логико-вероятностной модели распознавания образов

In this paper, we consider Bayes logical-and-probabilistic model of pattern recognition on a finite set of events, where event is defined as an assignment of some values from a subregion of the partition space described by a logical statement, to the given heterogeneous variables. Using this model, we found an expression for the expected probability of errors for the optimal Bayes decision function, obtain a posteriori estimations of the classifications probability for a given sample. These estimations are used for a construction and studying of the logical decision function.