To the Question of Increasing the Reliability of Measurements on the Basis of Application of Bayesian Approach

The proposed paper is devoted to analyzing the possibilities of Bayesian methods application for algorithmic support of measurement - information devices of complex technical systems. Six types of distribution laws are considered: normal law, uniform law, Poisson's law, exponential law, Weibull's law and Pareto's law. Algorithms and results of the construction of point estimates and interval estimates for the parameters of these laws are presented. The effectiveness of the methods is demonstrated on three examples showing that the confidence interval span can be reduced by a factor of two or more compared to the maximum likelihood method. Thus, the possibility of a multiple increase in the accuracy of measurements is shown. Finite formulas for transition from a priori distribution law to a posteriori distribution law are obtained for the specified distribution laws. The obtained formulas can be applied repeatedly as new relevant information is received by the measurement-information devices. Due to their simplicity, the formulas can be easily realized on a minicomputer and single-board computers. The described methods can be used as a basis for control and monitoring of complex controlled technical systems equipped with measurement information devices.