An Omnibus Non-Parametric Test of Equality in Distribution for Unknown Functions
Abstract
We present a novel family of non-parametric omnibus tests of the hypothesis that two unknown but estimable functions are equal in distribution when applied to the observed data structure. We developed these tests, which represent a generalization of the maximum mean discrepancy tests described by Gretton and colleagues, using recent developments from the higher order pathwise differentiability literature. Despite their complex derivation, the associated test statistics can be expressed quite simply as U-statistics. We study the asymptotic behaviour of the proposed tests under the null hypothesis and under both fixed and local alternatives. We provide examples to which our tests can be applied and show that they perform well in a simulation study. As an important special case, our proposed tests can be used to determine whether an unknown function, such as the conditional average treatment effect, is equal to zero almost surely.
Document Details
- Document Type
- Pub Defense Publication
- Publication Date
- Nov 02, 2018
- Source ID
- 10.1111/rssb.12299
Entities
People
- Alex Luedtke
- Marco Carone
- Mark J. Van Der Laan
Organizations
- National Institutes of Health
- United States Department of Defense
- University of California, Berkeley
- University of Washington