GOODNESS-OF-FIT TEST BASED ON BINOMIAL PROBABILITY DISTRIBUTION
E. L. Kuleshov, K. A. Petrov, T. S. Kirillova, R. A. Khaliullin
Far Eastern Federal University, 690950, Vladivostok, ul. Sukhanov, 8
Keywords: критерий согласия, интервальная оценка, закон распределения вероятностей, уровень значимости, goodness-of-fit test, interval estimate, probability distribution law, significance level
Subsection: ANALYSIS AND SYNTHESIS OF SIGNALS AND IMAGES
Abstract
This paper describes the goodness-of-fit test based on binomial probability distribution, which reduces to a sequence of two-sided hypothesis test for the value of the probability distribution function with different values of its argument. It is shown that each element of this sequence is unbiased locally by the most powerful test. This paper proposes an algorithm for calculating the significance level, free of probability distributions. The quality of this criterion is tested by numerical modeling.
|
