Eva Kotlebová, Faculty of Economic Informatics, University of Economics in Bratislava, Slovak Republic
Ivan Láska, TREXIMA Bratislava, Slovak Republic

Pages: 3 – 17

Abstract

The article deals with the point estimation of binomial proportion. In comparison withthe classical approach, bayesian approach takes into account another source ofinformation about the problem. In bayesian statistics, the conjugated family binomial/beta is used for the proportion´s estimation.
In the article, the mean square error was used as the estimation´s criterion of quality.
We proved, that if prior parameters´ values are appropriately determined, there exists interval, within which the bayesian point estimation has smaller mean square error than the classical one. We designed algorithm for evaluating the bayesian point estimation.