Information geometry of reversible Markov chains

DOI10.1007/S41884-021-00061-7zbMATH Open1479.60153arXiv2106.05669OpenAlexW3216853203MaRDI QIDQ2064255FDOQ2064255

Authors: Geoffrey Wolfer, Shun Watanabe 

Publication date: 5 January 2022

Published in: Information Geometry (Search for Journal in Brave)

Abstract: We analyze the information geometric structure of time reversibility for parametric families of irreducible transition kernels of Markov chains. We define and characterize reversible exponential families of Markov kernels, and show that irreducible and reversible Markov kernels form both a mixture family and, perhaps surprisingly, an exponential family in the set of all stochastic kernels. We propose a parametrization of the entire manifold of reversible kernels, and inspect reversible geodesics. We define information projections onto the reversible manifold, and derive closed-form expressions for the e-projection and m-projection, along with Pythagorean identities with respect to information divergence, leading to some new notion of reversiblization of Markov kernels. We show the family of edge measures pertaining to irreducible and reversible kernels also forms an exponential family among distributions over pairs. We further explore geometric properties of the reversible family, by comparing them with other remarkable families of stochastic matrices. Finally, we show that reversible kernels are, in a sense we define, the minimal exponential family generated by the m-family of symmetric kernels, and the smallest mixture family that comprises the e-family of memoryless kernels.

Full work available at URL: https://arxiv.org/abs/2106.05669

Recommendations

• Geodesic convexity of the relative entropy in reversible Markov chains
• Information geometry approach to parameter estimation in Markov chains
• The algebra of reversible Markov chains
• Geometric Markov chains
• Estimation in reversible Markov chains

zbMATH Keywords

exponential familyreversible Markov chainirreducible Markov chaininformation projectionmixture family

Mathematics Subject Classification ID

Random matrices (algebraic aspects) (15B52) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10)

Cites Work

• Markov chains and mixing times. With a chapter on ``Coupling from the past by James G. Propp and David B. Wilson.
• Asymptotic evaluation of certain markov process expectations for large time, I
• Handbook of Markov Chain Monte Carlo
• Title not available (Why is that?)
• Title not available (Why is that?)
• Zur Umkehrbarkeit der statistischen Naturgesetze
• Title not available (Why is that?)
• Bayesian analysis for reversible Markov chains
• Title not available (Why is that?)
• A Measure of Asymptotic Efficiency for Tests of a Hypothesis Based on the sum of Observations
• A Convexity Property in the Theory of Random Variables Defined on a Finite Markov Chain
• Conditional exponential families and a representation theorem for asymptotic inference
• On exponential families of Markov processes
• Explicit limit results for minimal sufficient statistics and maximum likelihood estimators in some Markov processes: Exponential families approach
• Finite-length analysis on tail probability for Markov chain and application to simple hypothesis testing
• Information geometry approach to parameter estimation in Markov chains
• Title not available (Why is that?)
• On Sequential Maximum Likelihood Estimation for Exponential Families of Stochastic Processes
• Large Sample Inference for Markovian Exponential Families with Application to Branching Processes with Immigration
• On the converse theorem in statistical hypothesis testing for Markov chains
• Maximum entropy and conditional probability
• On Large Deviations from the Invariant Measure
• Eigenvalue bounds on convergence to stationarity for nonreversible Markov chains, with an application to the exclusion process
• Zur Theorie der Markoffschen Ketten
• A. N. Kolmogorov - the founder of the theory of reversible Markov processes
• Conditional limit theorems under Markov conditioning
• Reversibility and stochastic networks. With a new preface
• Local equivalence problem in hidden Markov model
• Invariant measures of time-reversible Markov chains
• Exponential Families of Stochastic Processes: A Unifying Semimartingale Approach
• The algebra of reversible Markov chains
• Mixing time estimation in reversible Markov chains from a single sample path
• Doubly autoparallel structure on the probability simplex
• Computing the nearest reversible Markov chain.

Cited In (6)

• Geometric reduction for identity testing of reversible Markov chains
• Information geometry approach to parameter estimation in Markov chains
• Geometric aspects of data-processing of Markov chains
• Reversible Markov kernels and involutions on product spaces
• Information criteria for quantifying loss of reversibility in parallelized KMC
• Improved estimation of relaxation time in nonreversible Markov chains

This page was built for publication: Information geometry of reversible Markov chains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2064255)