Power in High-Dimensional Testing Problems

TL;DR

This paper analyzes the applicability of Fan et al.'s power enhancement method in high-dimensional testing, identifying conditions under which it can improve test power as dimensionality grows with sample size.

Contribution

It provides theoretical conditions for when the power enhancement principle can be applied in high-dimensional settings with increasing parameter space.

Findings

In fixed-dimensional regimes, some tests cannot be improved.

When dimensionality increases slowly with sample size, tests can often be improved.

Sufficient conditions are given for the applicability of the power enhancement method.

Abstract

Fan et al. (2015) recently introduced a remarkable method for increasing asymptotic power of tests in high-dimensional testing problems. If applicable to a given test, their power enhancement principle leads to an improved test that has the same asymptotic size, uniformly non-inferior asymptotic power, and is consistent against a strictly broader range of alternatives than the initially given test. We study under which conditions this method can be applied and show the following: In asymptotic regimes where the dimensionality of the parameter space is fixed as sample size increases, there often exist tests that can not be further improved with the power enhancement principle. However, when the dimensionality of the parameter space increases sufficiently slowly with sample size and a marginal local asymptotic normality (LAN) condition is satisfied, every test with asymptotic size smaller than one can be improved with the power enhancement principle. While the marginal LAN condition alone does not allow one to extend the latter statement to all rates at which the dimensionality increases with sample size, we give sufficient conditions under which this is the case.