Title: Two-sample smooth test for the equality of distributions for dependent data
Authors: Eric Beutner - Vrije Universiteit Amsterdam (Netherlands) [presenting]
Abstract: The two-sample problem is considered, i.e. testing whether two cumulative distribution functions are equal. Various proposals have been made to construct test statistics for this testing problem. These proposals include kernel-based methods, characteristic function type approaches, Cramer-von Mises type statistics, energy statistics and modifications of Neyman's smooth test. Common to these approaches is that they assume the two samples to be independent. We focus on modifications of Neyman's smooth test and extend them to dependent data. Bootstrapping the test statistics is also considered.