Multinomial goodness-of-fit based on U-statistics: high-dimensional asymptotic and minimax optimality

DOI10.1016/J.JSPI.2019.06.005zbMATH Open1432.62129arXiv1812.08924OpenAlexW2905593795MaRDI QIDQ2301048FDOQ2301048

Authors: Ilmun Kim 

Publication date: 28 February 2020

Published in: Journal of Statistical Planning and Inference (Search for Journal in Brave)

Abstract: We consider multinomial goodness-of-fit tests in the high-dimensional regime where the number of bins increases with the sample size. In this regime, Pearson's chi-squared test can suffer from low power due to the substantial bias as well as high variance of its statistic. To resolve these issues, we introduce a family of U-statistic for multinomial goodness-of-fit and study their asymptotic behaviors in high-dimensions. Specifically, we establish conditions under which the considered U-statistic is asymptotically Poisson or Gaussian, and investigate its power function under each asymptotic regime. Furthermore, we introduce a class of weights for the U-statistic that results in minimax rate optimal tests.

Full work available at URL: https://arxiv.org/abs/1812.08924

Recommendations

• Asymptotic approximations for the distributions of multinomial goodness- of-fit statistics
• Asymptotic distributions for goodness-of-fit statistics in a sequence of multinomial models
• Asymptotic approximations for the distributions of the multinomial goodness-of-fit statistics under local alternatives
• On the asymptotic properties of a certain class of goodness-of-fit tests associated with multinomial distributions
• A general family of limited information goodness-of-fit statistics for multinomial data
• On the intermediate asymptotic efficiency of goodness-of-fit tests in multinomial distributions
• Improvement of approximations for the distributions of multinomial goodness-of-fit statistics under nonlocal alternatives
• scientific article; zbMATH DE number 4166374

zbMATH Keywords

Poisson approximationmartingale central limit theoremmixture weightsparse multinomial distribution

Mathematics Subject Classification ID

Nonparametric hypothesis testing (62G10) Asymptotic properties of nonparametric inference (62G20)

Cites Work

• Testing Statistical Hypotheses
• Univariate Discrete Distributions
• Central limit theorem for integrated square error of multivariate nonparametric density estimators
• Hypothesis testing for densities and high-dimensional multinomials: sharp local minimax rates
• Hypothesis testing for high-dimensional multinomials: a selective review
• A two-sample test for high-dimensional data with applications to gene-set testing
• Title not available (Why is that?)
• Title not available (Why is that?)
• Mathematical Statistics
• Two moments suffice for Poisson approximations: The Chen-Stein method
• Central limit theorems for multinomial sums
• Asymptotic Distribution of The $\chi ^2 $ Criterion when the Number of Observations and Number of Groups Increase Simultaneously
• Title not available (Why is that?)
• Double asymptotics for the chi-square statistic
• Asymptotic normality and efficiency for certain goodness-of-fit tests
• Title not available (Why is that?)
• A Warning on the Use of Chi-Squared Statistics With Frequency Tables With Small Expected Cell Counts
• The matching, birthday and the strong birthday problem: a contemporary review
• Goodness-of-Fit Tests for Large Sparse Multinomial Distributions
• Chi-squared goodness of fit tests with applications
• The log-likelihood ratio for sparse multinomial mixtures
• An automatic inequality prover and instance optimal identity testing
• Collision-based Testers are Optimal for Uniformity and Closeness
• Geometry of goodness-of-fit testing in high dimensional low sample size modelling

Cited In (5)

• Asymptotically independent U-statistics in high-dimensional testing
• Poisson limit theorems for the Cressie-Read statistics
• Dimension-agnostic inference using cross U-statistics
• Geometry of goodness-of-fit testing in high dimensional low sample size modelling
• Testing for practically significant dependencies in high dimensions via bootstrapping maxima of \(U\)-statistics

This page was built for publication: Multinomial goodness-of-fit based on \(U\)-statistics: high-dimensional asymptotic and minimax optimality

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2301048)