On the minimal penalty for Markov order estimation

DOI10.1007/S00440-010-0290-YMaRDI QIDQ718903zbMATH OpenOpenAlexFDO

Authors Ramon van Handel

Publication date 27 September 2011

Published in Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/0908.3666

zbMATH Keywords

empirical process theorylarge-scale typicalityMarkov chainsmartingale inequalitiesuniform law of iterated logarithm

Mathematics Subject Classification ID

Inequalities; stochastic orderings (60E15) Strong limit theorems (60F15) Martingales with discrete parameter (60G42) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Markov processes: estimation; hidden Markov models (62M05)

Abstract: We show that large-scale typicality of Markov sample paths implies that the likelihood ratio statistic satisfies a law of iterated logarithm uniformly to the same scale. As a consequence, the penalized likelihood Markov order estimator is strongly consistent for penalties growing as slowly as log log n when an upper bound is imposed on the order which may grow as rapidly as log n. Our method of proof, using techniques from empirical process theory, does not rely on the explicit expression for the maximum likelihood estimator in the Markov case and could therefore be applicable in other settings.

Recommendations

• Large-scale typicality of Markov sample paths and consistency of MDL order estimators
• scientific article; zbMATH DE number 440531
• The consistency of the BIC Markov order estimator.
• Consistency of the BIC order estimator
• The optimal error exponent for Markov order estimation

Cites work

• Concentration inequalities and model selection. Ecole d'Eté de Probabilités de Saint-Flour XXXIII -- 2003.
• Context tree estimation for not necessarily finite memory processes, via BIC and MDL
• scientific article; zbMATH DE number 1420699 (Why is no real title available?)
• Large-scale typicality of Markov sample paths and consistency of MDL order estimators
• Strongly consistent code-based identification and order estimation for constrained finite-state model classes
• The consistency of the BIC Markov order estimator.
• Weak convergence and empirical processes. With applications to statistics

Cited in

(5)

• On universal algorithms for classifying and predicting stationary processes
• Divergence rates of Markov order estimators and their application to statistical estimation of stationary ergodic processes
• Large-scale typicality of Markov sample paths and consistency of MDL order estimators
• Personalized online ensemble machine learning with applications for dynamic data streams
• Model-based clustering in simple hypergraphs through a stochastic blockmodel

This page was built for publication: On the minimal penalty for Markov order estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q718903)