Convergence rate for predictive recursion estimation of finite mixtures

DOI10.1016/J.SPL.2011.10.023zbMATH Open1237.62044arXiv1106.4223OpenAlexW2063061677MaRDI QIDQ419247FDOQ419247

Authors: Ryan Martin 

Publication date: 18 May 2012

Published in: Statistics \& Probability Letters (Search for Journal in Brave)

Abstract: Predictive recursion (PR) is a fast stochastic algorithm for nonparametric estimation of mixing distributions in mixture models. It is known that the PR estimates of both the mixing and mixture densities are consistent under fairly mild conditions, but currently very little is known about the rate of convergence. Here I first investigate asymptotic convergence properties of the PR estimate under model misspecification in the special case of finite mixtures with known support. Tools from stochastic approximation theory are used to prove that the PR estimates converge, to the best Kullback--Leibler approximation, at a nearly root-$n$ rate. When the support is unknown, PR can be used to construct an objective function which, when optimized, yields an estimate the support. I apply the known-support results to derive a rate of convergence for this modified PR estimate in the unknown support case, which compares favorably to known optimal rates.

Full work available at URL: https://arxiv.org/abs/1106.4223

Recommendations

• Asymptotic properties of predictive recursion: robustness and rate of convergence
• A survey of nonparametric mixing density estimation via the predictive recursion algorithm
• Semiparametric inference in mixture models with predictive recursion marginal likelihood
• A PRticle filter algorithm for nonparametric estimation of multivariate mixing distributions
• Consistency of a recursive estimate of mixing distributions

zbMATH Keywords

density estimationKullback-Leibler divergencestochastic approximationLyapunov function

Mathematics Subject Classification ID

Density estimation (62G07) Asymptotic properties of nonparametric inference (62G20) Stochastic approximation (62L20)

Cites Work

• Bayesian Density Estimation and Inference Using Mixtures
• Mixture models: theory, geometry and applications
• Bayesian Model Selection in Finite Mixtures by Marginal Density Decompositions
• Entropies and rates of convergence for maximum likelihood and Bayes estimation for mixtures of normal densities.
• Title not available (Why is that?)
• A Stochastic Approximation Method
• Density Estimation With Confidence Sets Exemplified by Superclusters and Voids in the Galaxies
• Title not available (Why is that?)
• Stochastic approximation and its applications
• Stability by Liapunov's direct method. With applications
• Geometric bounds for eigenvalues of Markov chains
• Optimal rate of convergence for finite mixture models
• Order selection in finite mixture models with a nonsmooth penalty
• Rates of convergence for the Gaussian mixture sieve.
• Robust estimation of mixture complexity for count data
• Robust inference for finite Poisson mixtures
• Asymptotic properties of predictive recursion: robustness and rate of convergence
• A nonparametric empirical Bayes framework for large-scale multiple testing
• Semiparametric inference in mixture models with predictive recursion marginal likelihood
• On a nonparametric recursive estimator of the mixing distribution
• Title not available (Why is that?)
• Convergence and consistency of Newton's algorithm for estimating mixing distribution
• Consistency of a recursive estimate of mixing distributions
• Stochastic approximation and Newton's estimate of a mixing distribution

Cited In (8)

• An Approximate Bayesian Marginal Likelihood Approach for Estimating Finite Mixtures
• Strong identifiability and optimal minimax rates for finite mixture estimation
• Asymptotic properties of predictive recursion: robustness and rate of convergence
• A survey of nonparametric mixing density estimation via the predictive recursion algorithm
• Asymptotically optimal nonparametric empirical Bayes via predictive recursion
• Semiparametric inference in mixture models with predictive recursion marginal likelihood
• A PRticle filter algorithm for nonparametric estimation of multivariate mixing distributions
• Optimal Bayesian estimation of Gaussian mixtures with growing number of components

This page was built for publication: Convergence rate for predictive recursion estimation of finite mixtures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q419247)