Some remarks on MCMC estimation of spectra of integral operators

DOI10.3150/14-BEJ635zbMATH Open1350.60027arXiv1311.7566OpenAlexW1954519420MaRDI QIDQ888474FDOQ888474

Authors: Radosław Adamczak, Witold Bednorz 

Publication date: 30 October 2015

Published in: Bernoulli (Search for Journal in Brave)

Abstract: We prove a law of large numbers for empirical approximations of the spectrum of a kernel integral operator by the spectrum of random matrices based on a sample drawn from a Markov chain, which complements the results by V. Koltchinskii and E. Gin'{e} for i.i.d. sequences. In a special case of Mercer's kernels and geometrically ergodic chains, we also provide exponential inequalities, quantifying the speed of convergence.

Full work available at URL: https://arxiv.org/abs/1311.7566

Recommendations

• Random matrix approximation of spectra of integral operators
• Estimating the spectral gap of a trace-class Markov operator
• Spectral bounds for certain two-factor non-reversible MCMC algorithms
• The spectrum of the independent Metropolis-Hastings algorithm
• MC's for MCMC'ists

zbMATH Keywords

law of large numbersrandom matricesintegral operatorsspectrakernel operatorsMarkov chain Monte Carlo algorithmsempirical approximations

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Random matrices (probabilistic aspects) (60B20) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Strong limit theorems (60F15) Integral operators (45P05) Integral operators (47G10)

Cites Work

• Title not available (Why is that?)
• Consistency of spectral clustering
• Markov chains and stochastic stability
• General state space Markov chains and MCMC algorithms
• General Irreducible Markov Chains and Non-Negative Operators
• A tail inequality for suprema of unbounded empirical processes with applications to Markov chains
• Limit theorems for functionals of ergodic Markov chains with general state space
• On learning with integral operators
• Renewal theory and computable convergence rates for geometrically erdgodic Markov chains
• Random matrix approximation of spectra of integral operators
• A splitting technique for Harris recurrent Markov chains
• Mercer theorem for RKHS on noncompact sets
• The Hoffman-Wielandt inequality in infinite dimensions
• On nonlinear Markov chain Monte Carlo
• On the Eigenspectrum of the Gram Matrix and the Generalization Error of Kernel-PCA
• Title not available (Why is that?)
• Geometry on probability spaces
• A New Approach to the Limit Theory of Recurrent Markov Chains
• Strong laws for 𝐿- and 𝑢-statistics
• The law of large numbers for \(U\)-statistics under absolute regularity
• On singular values of matrices with independent rows
• Practical drift conditions for subgeometric rates of convergence.
• Bounds on regeneration times and limit theorems for subgeometric Markov chains
• Mercer’s Theorem, Feature Maps, and Smoothing
• The Kendall theorem and its application to the geometric ergodicity of Markov chains
• A renewal approach to Markovian \(U\)-statistics
• Title not available (Why is that?)
• Learning Theory

Cited In (8)

• Consistent estimation of the spectrum of trace class data augmentation algorithms
• On the kernel of the covariance operator for Markov semigroups. I.
• Computing the asymptotic distribution of second-order \(U\)- and \(V\)-statistics
• Random discretization of the finite Fourier transform and related kernel random matrices
• From random matrices to long range dependence
• Trace class Markov chains for the normal-gamma Bayesian shrinkage model
• Trace-class Monte Carlo Markov chains for Bayesian multivariate linear regression with non-Gaussian errors
• Estimating the spectral gap of a trace-class Markov operator

This page was built for publication: Some remarks on MCMC estimation of spectra of integral operators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q888474)