Game-theoretic statistics and safe anytime-valid inference

DOI10.1214/23-STS894arXiv2210.01948MaRDI QIDQ6145149FDOQ6145149

Authors: Aaditya Ramdas, Peter D. Grünwald, Vladimir Vovk, Glenn Shafer 

Publication date: 30 January 2024

Published in: Statistical Science (Search for Journal in Brave)

Abstract: Safe anytime-valid inference (SAVI) provides measures of statistical evidence and certainty -- e-processes for testing and confidence sequences for estimation -- that remain valid at all stopping times, accommodating continuous monitoring and analysis of accumulating data and optional stopping or continuation for any reason. These measures crucially rely on test martingales, which are nonnegative martingales starting at one. Since a test martingale is the wealth process of a player in a betting game, SAVI centrally employs game-theoretic intuition, language and mathematics. We summarize the SAVI goals and philosophy, and report recent advances in testing composite hypotheses and estimating functionals in nonparametric settings.

Full work available at URL: https://arxiv.org/abs/2210.01948

zbMATH Keywords

e-processoptional stoppingconfidence sequencetest martingalesuniversal inferenceVille's inequalityreverse information projectionnonparametric composite hypothesis testing

Mathematics Subject Classification ID

Statistics (62-XX)

Cites Work

• Title not available (Why is that?)
• On confidence sequences
• The control of the false discovery rate in multiple testing under dependency.
• Time-uniform, nonparametric, nonasymptotic confidence sequences
• Present Position and Potential Developments: Some Personal Views: Statistical Theory: The Prequential Approach
• Always Valid Inference: Continuous Monitoring of A/B Tests
• Valid sequential inference on probability forecast performance
• A Sequential Non-Parametric Multivariate Two-Sample Test
• CONFIDENCE SEQUENCES FOR MEAN, VARIANCE, AND MEDIAN
• Statistical Methods Related to the Law of the Iterated Logarithm
• Title not available (Why is that?)
• Sequential Tests of Statistical Hypotheses
• Likelihood, Replicability and Robbins' Confidence Sequences
• Exact anytime-valid confidence intervals for contingency tables and beyond
• Test martingales, Bayes factors and \(p\)-values
• E-values: calibration, combination and applications
• Challenging the empirical mean and empirical variance: a deviation study
• Panning for Gold: ‘Model-X’ Knockoffs for High Dimensional Controlled Variable Selection
• Title not available (Why is that?)
• A general class of exponential inequalities for martingales and ratios
• Some aspects of the sequential design of experiments
• Title not available (Why is that?)
• Student's t-Test Under Symmetry Conditions
• Probability and finance. It's only a game!
• Universal coding, information, prediction, and estimation
• The minimum description length principle in coding and modeling
• Bayes factors and marginal distributions in invariant situations
• False Discovery Rate–Adjusted Multiple Confidence Intervals for Selected Parameters
• Fixed-Sample-Size Analysis of Sequential Observations
• Exponential inequalities for martingales with applications
• Title not available (Why is that?)
• The expected sample size of some tests of power one
• Title not available (Why is that?)
• Insuring against loss of evidence in game-theoretic probability
• Interactive martingale tests for the global null
• Game-theoretic foundations for probability and finance
• Algorithmic Learning in a Random World
• SOME NONPARAMETRIC SEQUENTIAL TESTS WITH POWER ONE
• A TCHEBYCHEFF-LIKE INEQUALITY FOR STOCHASTIC PROCESSES
• Exponential inequalities for sums of weakly dependent variables
• ON A SEQUENTIAL t-TEST
• Title not available (Why is that?)
• Sequential estimation of quantiles with applications to A/B testing and best-arm identification
• Time-uniform Chernoff bounds via nonnegative supermartingales
• Catoni-style confidence sequences for heavy-tailed mean estimation
• Testing exchangeability: fork-convexity, supermartingales and e-processes
• Testing randomness online
• Universal inference
• Title not available (Why is that?)
• Gaussian universal likelihood ratio testing

Cited In (3)

• E-statistics, group invariance and anytime-valid testing
• On the existence of powerful p-values and e-values for composite hypotheses
• Game-theoretic statistical inference: optional sampling, universal inference, and multiple testing based on e-values. Abstracts from the workshop held May 5--10, 2024

This page was built for publication: Game-theoretic statistics and safe anytime-valid inference

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6145149)