INTERVAL ESTIMATION OF THE PROBABILITY DISTRIBUTION FUNCTION
E. L. Kuleshov
Far Eastern Federal University, ul. Sukhanova 8, Vladivostok, 690950 Russia
Keywords: interval estimate, probability distribution function, confidence factor, goodnessoffit test, statistical hypothesis
Abstract
The Moivre–Laplace asymptotics is used to construct an interval estimate of the probability distribution function that is an interval with random boundaries, which covers the true value of the distribution function with a given confidence factor. It is shown that the use of the asymptotic instead of a binomial probability distribution results in an error whose value is tolerable for small sampling sizes and monotonically reduces with decreases sampling size.
