Simultaneous Inference for Non-Stationary Random Fields, with Application to Gridded Data Analysis

TL;DR

This paper develops a nonparametric framework for simultaneous inference on complex non-stationary, non-Gaussian random fields, using a flexible modeling approach and a novel bootstrap method, with demonstrated effectiveness on real data.

Contribution

It introduces a nonparametric modeling approach for non-stationary, non-Gaussian random fields and proposes a computationally efficient bootstrap method for simultaneous inference.

Findings

The proposed bootstrap method performs well in simulations.

The approach is applicable to real-world gridded data.

Theoretical results support the validity of the inference.

Abstract

Current statistics literature on statistical inference of random fields typically assumes that the fields are stationary or focuses on models of non-stationary Gaussian fields with parametric/semiparametric covariance families, which may not be sufficiently flexible to tackle complex modern-era random field data. This paper performs simultaneous nonparametric statistical inference for a general class of non-stationary and non-Gaussian random fields by modeling the fields as nonlinear systems with location-dependent transformations of an underlying `shift random field'. Asymptotic results, including concentration inequalities and Gaussian approximation theorems for high dimensional sparse linear forms of the random field, are derived. A computationally efficient locally weighted multiplier bootstrap algorithm is proposed and theoretically verified as a unified tool for the simultaneous inference of the aforementioned non-stationary non-Gaussian random field. Simulations and real-life data examples demonstrate good performances and broad applications of the proposed algorithm.