Multidimensional multiscale scanning in exponential families: limit theory and statistical consequences

DOI10.1214/18-AOS1806zbMATH Open1450.60028arXiv1802.07995WikidataQ115517773 ScholiaQ115517773MaRDI QIDQ2196187FDOQ2196187

Authors: Claudia König, Axel Munk, Frank Werner 

Publication date: 28 August 2020

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: We consider the problem of finding anomalies in a $d$-dimensional field of independent random variables ${Y_i}_{iinleft{1,\mathrm{...},night}^d}$, each distributed according to a one-dimensional natural exponential family $mathcalF=left{F_{h}etaight}_{hetainTheta}$. Given some baseline parameter $het{a}_{0}inTheta$, the field is scanned using local likelihood ratio tests to detect from a (large) given system of regions $mathcalR$ those regions $Rsubsetleft{1,\mathrm{...},night}^d$ with $het{a}_{i}eqhet{a}_0$ for some $iinR$. We provide a unified methodology which controls the overall family wise error (FWER) to make a wrong detection at a given error rate. Fundamental to our method is a Gaussian approximation of the distribution of the underlying multiscale test statistic with explicit rate of convergence. From this, we obtain a weak limit theorem which can be seen as a generalized weak invariance principle to non identically distributed data and is of independent interest. Furthermore, we give an asymptotic expansion of the procedures power, which yields minimax optimality in case of Gaussian observations.

Full work available at URL: https://arxiv.org/abs/1802.07995

Recommendations

• Multiscale scanning in higher dimensions: limit theory, statistical consequences and an application in STED microscopy
• Tail probabilities for the null distribution of scanning statistics
• Large-deviation approximations to the distribution of scan statistics
• Exact asymptotics for the scan statistic and fast alternatives
• Multiple Window Discrete Scan Statistics

zbMATH Keywords

exponential familiesweak limitinvariance principlefamilywise error ratescan statisticmultiscale testing

Mathematics Subject Classification ID

Parametric hypothesis testing (62F03) Multivariate distribution of statistics (62H10) Random fields (60G60) Paired and multiple comparisons; multiple testing (62J15) Functional limit theorems; invariance principles (60F17)

Cites Work

• Weak convergence and empirical processes. With applications to statistics
• Title not available (Why is that?)
• Testing Statistical Hypotheses
• FDR-control in multiscale change-point segmentation
• On excursion sets, tube formulas and maxima of random fields.
• Multiscale testing of qualitative hypotheses
• The control of the false discovery rate in multiple testing under dependency.
• Detection of an anomalous cluster in a network
• Maximum likelihood estimation in a class of nonregular cases
• Title not available (Why is that?)
• An approximation of partial sums of independent RV's, and the sample DF. II
• Distribution-free detection of structured anomalies: permutation and rank-based scans
• Simultaneous Statistical Inference
• Rectangular Confidence Regions for the Means of Multivariate Normal Distributions
• Multiscale Change Point Inference
• Multiscale inference about a density
• Title not available (Why is that?)
• Title not available (Why is that?)
• Using the generalized likelihood ratio statistic for sequential detection of a change-point
• Multiple testing of local maxima for detection of peaks in random fields
• Multiple testing of local maxima for detection of peaks in 1D
• Gaussian approximation of suprema of empirical processes
• Detection of a sparse submatrix of a high-dimensional noisy matrix
• Tail probabilities for the null distribution of scanning statistics
• Optimal and fast detection of spatial clusters with scan statistics
• Optimal detection of a jump in the intensity of a Poisson process or in a density with likelihood ratio statistics
• Combinatorial methods in density estimation
• Extremes of the standardized Gaussian noise
• Detecting Sparse Signals in Random Fields, With an Application to Brain Mapping
• Near-Optimal Detection of Geometric Objects by Fast Multiscale Methods
• Multiple window and cluster size scan procedures
• Maxima of partial sums indexed by geometrical structures
• Scan statistics on Poisson random fields with applications in genomics
• Exact asymptotics for the scan statistic and fast alternatives
• A martingale approach to scan statistics
• Estimation for the distribution of two-dimensional discrete scan statistics
• Strong approximation for multivariate empirical and related processes, via KMT constructions
• Poisson approximation for two scan statistics with rates of convergence
• Multiscale methods for shape constraints in deconvolution: confidence statements for qualitative features
• Shao's theorem on the maximum of standardized random walk increments for multidimensional arrays
• Strong approximation for set-indexed partial-sum processes, via KMT constructions. II
• Multiscale scanning in inverse problems
• Approximation and simulation of the distributions of scan statistics for Poisson processes in higher dimensions
• Statistical detection of defects in radiographic images in nondestructive testing
• Straight to the Source: Detecting Aggregate Objects in Astronomical Images With Proper Error Control
• The maximum of all ratchet scanning process over a Poisson random field

Cited In (7)

• Scanning and Sequential Decision Making for Multidimensional Data–Part I: The Noiseless Case
• Sharp adaptive and pathwise stable similarity testing for scalar ergodic diffusions
• Optimal inference with a multidimensional multiscale statistic
• Tail bounds for empirically standardized sums
• On the asymptotic distribution of the scan statistic for empirical distributions
• Multiscale scanning in higher dimensions: limit theory, statistical consequences and an application in STED microscopy
• Detecting structured signals in Ising models

Uses Software

• FDRSeg

This page was built for publication: Multidimensional multiscale scanning in exponential families: limit theory and statistical consequences

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2196187)