Statistické rozhodování a jeho souvislost se ztrátovou funkcí

Abstract

Bayesian statistics is a modern discipline built on, at first look rather rudimentary, the Bayes theorem on conditional probability. Bayesian approach to statistics is intuitive but at the same time very deep and varied. The goal of this thesis was to provide a vertical slice of Bayesian Statistics. That is to introduce decision-making under uncertainty, loss functions, precautionary loss functions and numeric integration via Markov Chain Monte Carlo methods. Present simulated annealing as~a~method for numerical solution to optimalisation problems in decision-making.