Scientific journal
Modern high technologies
ISSN 1812-7320
"Перечень" ВАК
ИФ РИНЦ = 1,279

Personal portfolio
(submit article)

BAYESIAN APPROACH TO ACCOUNTING FOR EPISTEMIC UNCERTAINTY OF PARAMETERS OF PROBABILISTIC MODELS OF RISK ANALYSIS OF DECISIONS

Ptushkin A.I. 1, Reshetnikov D.V. 1, Shapovalov D.V. 1, Stepenko A.N. 1
1 Mozhaisky Military Space Academy
The problem of reducing the level of uncertainty of knowledge about the parameters of probabilistic failure models used in the risk analysis of decisions is considered. The definition of risk as one of the properties of a decision made in a situation with uncertain outcomes is proposed. In order to obtain additional quantitative knowledge about the parameters of probabilistic failure models, it is proposed to use the Bayesian approach. The General scheme of application of this approach is presented. It is demonstrated that due to the possibility of using widely available application programs contained in the Excel spreadsheet, the application of this approach has become characterized by simplicity and significantly less labor-intensive compared to General approaches to uncertainty research based on the Dempster-Schaefer proof theory or the concept of second-order probabilities (histograms). It is considered how in the process of operation of complex technical devices, using the proposed approach, it is possible to adjust the values of parameters of mathematical models of failures necessary for analyzing the risk of decisions made, using the example of two distribution laws: binomial and exponential. Comparisons of a priori and a posteriori probability densities of their parameters are shown, indicating a decrease in the epistemic uncertainty of a priori distributions when additional observations are taken into account. Recommendations are given for using the developed approach to analyze the risk of decision-making at the operational stage of the product life cycle.
epistemic uncertainty
risk
confidence interval
Bayesian approach
a priori and a posteriori probability distribution densities
conjugate distribution

Библиографическая ссылка

Птушкин А.И., Решетников Д.В., Шаповалов Д.В., Степенко А.Н. БАЙЕСОВСКИЙ ПОДХОД К УЧЕТУ ЭПИСТЕМИЧЕСКОЙ НЕОПРЕДЕЛЕННОСТИ ПАРАМЕТРОВ ВЕРОЯТНОСТНЫХ МОДЕЛЕЙ АНАЛИЗА РИСКА ПРИНИМАЕМЫХ РЕШЕНИЙ // Современные наукоемкие технологии. 2020. № 11-1. С. 67-72;
URL: https://top-technologies.ru/en/article/view?id=38300 (дата обращения: 22.06.2026).