Abstract:Heteroskedasticity testing in nonparametric regression is a classic statistical problem with important practical applications, yet fundamental limits are unknown. Adopting a minimax perspective, this article considers the testing problem in the context of an $\alpha$-Hölder mean and a $\beta$-Hölder variance function. For $\alpha > 0$ and $\beta \in (0, 1/2)$, the sharp minimax separation rate $n^{-4\alpha} + n^{-4\beta/(4\beta+1)} + n^{-2\beta}$ is established. To achieve the minimax separation rate, a kernel-based statistic using first-order squared differences is developed. Notably, the statistic estimates a proxy rather than a natural quadratic functional (the squared distance between the variance function and its best $L^2$ approximation by a constant) suggested in previous work. The setting where no smoothness is assumed on the variance function is also studied; the variance profile across the design points can be arbitrary. Despite the lack of structure, consistent testing turns out to still be possible by using the Gaussian character of the noise, and the minimax rate is shown to be $n^{-4\alpha} + n^{-1/2}$. Exploiting noise information happens to be a fundamental necessity as consistent testing is impossible if nothing more than zero mean and unit variance is known about the noise distribution. Furthermore, in the setting where the variance function is $\beta$-Hölder but heteroskedasticity is measured only with respect to the design points, the minimax separation rate is shown to be $n^{-4\alpha} + n^{-\left((1/2) \vee (4\beta/(4\beta+1))\right)}$ when the noise is Gaussian and $n^{-4\alpha} + n^{-4\beta/(4\beta+1)} + n^{-2\beta}$ when the noise distribution is unknown.

Testing for homogeneity of several functional variables via multiple maximum variance discrepancy

Testing for Homogeneity with Kernel Fisher Discriminant Analysis

Testing distributional equality for functional random variables

Variable Selection in Maximum Mean Discrepancy for Interpretable Distribution Comparison

A uniform kernel trick for high-dimensional two-sample problems

Testing equality between several populations covariance operators

Testing high-dimensional multinomials with applications to text analysis

Two Sample Testing in High Dimension via Maximum Mean Discrepancy

A Kernel Method for the Two-Sample Problem

Homogeneity Test of Several High-Dimensional Covariance Matrices for Stationary Processes under Non-normality

Testing marginal homogeneity in Hilbert spaces with applications to stock market returns

A Multivariate Unimodality Test Harnessing the Dip Statistic of Mahalanobis Distances Over Random Projections

Optimal heteroskedasticity testing in nonparametric regression

Minimax Optimal Goodness-of-Fit Testing with Kernel Stein Discrepancy

Testing the equality of multivariate means when $p>n$ by combining the Hoteling and Simes tests

Two-Sample Test for Sparse High Dimensional Multinomial Distributions

Testing Homogeneity in A Multivariate Mixture Model

Double verification for two‐sample covariance matrices test

Asymptotic normality of a generalized maximum mean discrepancy estimator

A Global Homogeneity Test for High-Dimensional Linear Regression

Homogeneity Test of Several Covariance Matrices with High-Dimensional Data