On the global and componentwise rates of convergence of the EM algorithm

DOI10.1016/0024-3795(94)90363-8MaRDI QIDQ1322882zbMATH OpenOpenAlexFDO

Authors Donald B. Rubin, Xiao-Li Meng

Publication date 20 July 1995

Published in Linear Algebra and its Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/0024-3795(94)90363-8

zbMATH Keywords

convergenceEM algorithmincomplete-data problemsmaximum-likelihood estimates

Mathematics Subject Classification ID

Estimation in multivariate analysis (62H12) Inference from stochastic processes and spectral analysis (62M15) Probabilistic methods, stochastic differential equations (65C99)

Review text

For incomplete-data problems in statistics, the EM algorithm is a very popular one for finding maximum-likelihood estimates and posterior modes. It is simple to apply and to implement, and it is stable. Each iteration of EM consists of an expectation step and a maximization step. After recalling that the order of the EM algorithm is generally linear, the authors focus on the rate of convergence for linear iterations. By using the diagonalizability theorem, they describe how and when the componentwise rates differ and their relationships with the global rate. This study is nicely enlightened by an example of a standard contaminated normal model which illustrates that these phenomena are not necessarily pathological and can occur in useful statistical models.

Recommendations

• scientific article; zbMATH DE number 4032799
• On the rate of convergence of the ECM algorithm
• Simple and Globally Convergent Methods for Accelerating the Convergence of Any EM Algorithm
• On convergence properties of the Monte Carlo EM algorithm
• On the rate of convergence of the ECME algorithm

Cites work

• scientific article; zbMATH DE number 3885090 (Why is no real title available?)
• scientific article; zbMATH DE number 4088699 (Why is no real title available?)
• scientific article; zbMATH DE number 3181418 (Why is no real title available?)
• scientific article; zbMATH DE number 3522963 (Why is no real title available?)
• scientific article; zbMATH DE number 3567782 (Why is no real title available?)
• scientific article; zbMATH DE number 3802652 (Why is no real title available?)
• scientific article; zbMATH DE number 3415203 (Why is no real title available?)
• Maximum likelihood estimation via the ECM algorithm: A general framework
• On the convergence properties of the EM algorithm

Cited in

(31)

• Accelerating the convergence of the EM algorithm using the vector algorithm
• On the rate of convergence of the ECM algorithm
• A combined likelihood ratio/information ratio bootstrap technique for estimating the number of components in finite mixtures
• Global convergence and ascent property of a cyclic algorithm used for statistical analysis of crash data
• On spectral radii of matrices governing some numerical estimating algorithms with incomplete data
• Periodic step-size adaptation in second-order gradient descent for single-pass on-line structured learning
• On computing the largest fraction of missing information for the EM algorithm and the worst linear function for data augmentation.
• Fitting a Gaussian mixture model through the Gini index
• Finite mixture of Birnbaum-Saunders distributions using the k-bumps algorithm
• Model-based sparse coding beyond Gaussian independent model
• Global convergence of the EM algorithm for ARX models with uncertain communication channels
• Fixed points of the EM algorithm and nonnegative rank boundaries
• Robust maximum-likelihood estimation of multivariable dynamic systems
• Convergence of E-MS algorithm
• Maximum Likelihood Estimation of VARMA Models Using a State-Space EM Algorithm
• On Casting Random-Effects Models in a Survival Framework
• scientific article; zbMATH DE number 4032799 (Why is no real title available?)
• Theoretical and practical considerations on the convergence properties of the Fisher-EM algorithm
• scientific article; zbMATH DE number 739537 (Why is no real title available?)
• Convergence rate of the EM algorithm for SDEs with low regular drifts
• Statistical guarantees for the EM algorithm: from population to sample-based analysis
• On the Slow Convergence of EM and VBEM in Low-Noise Linear Models
• Regression modeling of censored data based on compound scale mixtures of normal distributions
• Acceleration of the EM algorithm using the vector epsilon algorithm
• Convergence of parameter estimation of a Gaussian mixture model minimizing the Gini index of dissimilarity
• Fast convergence of the expectation-maximization algorithm under a logarithmic Sobolev inequality
• Quasi-Newton Acceleration of EM and MM Algorithms via Broyden’s Method
• Modelling large dimensional datasets with Markov switching factor models
• Improving the vector \(\varepsilon\) acceleration for the EM algorithm using a re-starting procedure
• Cross-fertilizing strategies for better EM mountain climbing and DA field exploration: a graphical guide book
• Discussion: One-step sparse estimates in nonconcave penalized likelihood models: Who cares if it is a white cat or a black cat?

This page was built for publication: On the global and componentwise rates of convergence of the EM algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1322882)