Notes on the dimension dependence in high-dimensional central limit theorems for hyperrectangles
2020
Let
$$X_1,\ldots ,X_n$$
be independent centered random vectors in
$${\mathbb {R}}^d$$
. This paper shows that, even when d may grow with n, the probability
$$P(n^{-1/2}\sum _{i=1}^nX_i\in A)$$
can be approximated by its Gaussian analog uniformly in hyperrectangles A in
$${\mathbb {R}}^d$$
as
$$n\rightarrow \infty$$
under appropriate moment assumptions, as long as
$$(\log d)^5/n\rightarrow 0$$
. This improves a result of Chernozhukov et al. (Ann Probab 45:2309–2353, 2017) in terms of the dimension growth condition. When
$$n^{-1/2}\sum _{i=1}^nX_i$$
has a common factor across the components, this condition can be further improved to
$$(\log d)^3/n\rightarrow 0$$
. The corresponding bootstrap approximation results are also developed. These results serve as a theoretical foundation of simultaneous inference for high-dimensional models.
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
31
References
9
Citations
NaN
KQI