Detecting relevant changes in the spatiotemporal mean function

TL;DR

This paper develops asymptotically distribution-free tests for detecting significant changes in the mean function of spatiotemporal processes, focusing on changes exceeding a specified threshold without needing covariance estimation.

Contribution

It introduces new change detection tests that are distribution-free and do not require covariance estimation, applicable to fully functional and cumulative sum based approaches.

Findings

Tests are asymptotically distribution-free.

Finite sample properties are validated via simulations.

Methods effectively detect meaningful changes exceeding the threshold.

Abstract

For a spatiotemporal process $\{X_j(s,t) | ~s \in S~,~t \in T \}_{j =1, \ldots , n} $, where $S$ denotes the set of spatial locations and $T$ the time domain, we consider the problem of testing for a change in the sequence of mean functions. In contrast to most of the literature we are not interested in arbitrarily small changes, but only in changes with a norm exceeding a given threshold. Asymptotically distribution free tests are proposed, which do not require the estimation of the long-run spatiotemporal covariance structure. In particular we consider a fully functional approach and a test based on the cumulative sum paradigm, investigate the large sample properties of the corresponding test statistics and study their finite sample properties by means of simulation study.