Sequential Probability Ratio Test for Mean Based on Pseudo-Likelihood Function
Dorota Anna Pekasiewicz · University of Lodz Repository (University of Łódź) · 2009
Hypotheses about expected value of random variable can be verified by means of the parametric sequential probability ratio test in case of the known class of this variable's distribution. The problem with verification of such hypotheses occurs when we have no information about random variable distribution. Then, we have to apply non-parametric methods. The author of the paper proposes the application of pseudo-likelihood function instead of likelihood one in the statistic of sequential probability ratio test. Examples of application of the test based on the likelihood function ratio in selected kinds of distributions are presented together with the results of Monte Carlo analysis concerning properties of these tests.