A Numerical Analysis Procedure for Improving Performance of Hoeffding Probability Inequality

Abstract

Hoeffding probability inequality has been known as the best theoretical way for evaluating the upper bound of upper probability about the average of some random variables based on limited stochastic properties such as the expectation, variance and codomain in each random variable. In this study, based on Hoeffding probability inequality, we address a numerical analysis procedure for improving the performance to evaluate the upper bound of upper probability about the average of some random variables.