English

A Ball Divergence Based Measure For Conditional Independence Testing

Statistics Theory 2024-08-01 v1 Methodology Statistics Theory

Abstract

In this paper we introduce a new measure of conditional dependence between two random vectors X{\boldsymbol X} and Y{\boldsymbol Y} given another random vector Z\boldsymbol Z using the ball divergence. Our measure characterizes conditional independence and does not require any moment assumptions. We propose a consistent estimator of the measure using a kernel averaging technique and derive its asymptotic distribution. Using this statistic we construct two tests for conditional independence, one in the model-X{\boldsymbol X} framework and the other based on a novel local wild bootstrap algorithm. In the model-X{\boldsymbol X} framework, which assumes the knowledge of the distribution of XZ{\boldsymbol X}|{\boldsymbol Z}, applying the conditional randomization test we obtain a method that controls Type I error in finite samples and is asymptotically consistent, even if the distribution of XZ{\boldsymbol X}|{\boldsymbol Z} is incorrectly specified up to distance preserving transformations. More generally, in situations where XZ{\boldsymbol X}|{\boldsymbol Z} is unknown or hard to estimate, we design a double-bandwidth based local wild bootstrap algorithm that asymptotically controls both Type I error and power. We illustrate the advantage of our method, both in terms of Type I error and power, in a range of simulation settings and also in a real data example. A consequence of our theoretical results is a general framework for studying the asymptotic properties of a 2-sample conditional VV-statistic, which is of independent interest.

Keywords

Cite

@article{arxiv.2407.21456,
  title  = {A Ball Divergence Based Measure For Conditional Independence Testing},
  author = {Bilol Banerjee and Bhaswar B. Bhattacharya and Anil K. Ghosh},
  journal= {arXiv preprint arXiv:2407.21456},
  year   = {2024}
}
R2 v1 2026-06-28T17:59:06.836Z