UDC 510.644.4
Authors
Bautin Alexander G.
Student, Law faculty, Financial University under the Government of the Russian Federation, Moscow,
Russia
Scientific Supervisor – Zvyagin Leonid S., Candidate of Economic Sciences, Associate Professor,
Financial University, Moscow
Abstract
The Bayesian approach is often used in theoretical and experimental econometric research. It is especially popular in the processing of numerical information. The advantage of this method over the classical ones is a higher degree of accuracy of statistical conclusions in cases where the sample size is not very large. As a rule, such samples are used in econometric research. Interpretation of real parameters of the model is the main distinguishing feature of the Bayesian approach from classical methods, which claim that the parameters are not random, but their estimates, which are observation functions that include elements of randomness. In contrast, Bayesian methods state that the randomness of parameters is a property of the real world, and that each physical object is subject to constant random variations. Based on the Bayesian approach, estimates of these parameters are not random, so they should be calculated. An example of such estimates is the average value of a random variable. On the basis of the accepted hypothesis about the randomness of the model parameters, the Bayes theorem is applied. The main idea of the Bayesian approach is that by combining the a priori probability distribution density function of a set of parameters using Bayes' theorem, one can obtain the a posteriori distribution density function
Keywords
Bayesian approach, Uncertainty, Data analysis, Numerical sampling, Probability estimates
