ROBUST SEQUENTIAL TEST FOR HYPOTHESES ABOUT DISCRETE DISTRIBUTIONS IN THE PRESENCE OF 'OUTLIERS'

摘要

In this paper a robust sequential test is constructed for a discrete observation model, for which probabilities of first-type and second-type errors under Tukey-Huber type distortions do not exceed the specified values. No additional "commensurability" condition is imposed on the model, and test construction does not require cumbersome calculations. Theoretically studied properties of the constructed test are illustrated by numerical examples.