Existence and uniqueness of the Kronecker covariance MLE

DOI10.1214/21-AOS2052zbMATH Open1480.62096arXiv2003.06024OpenAlexW3213304819MaRDI QIDQ2054526FDOQ2054526

Authors: Mathias Drton, Satoshi Kuriki, Peter D. Hoff 

Publication date: 3 December 2021

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: In matrix-valued datasets the sampled matrices often exhibit correlations among both their rows and their columns. A useful and parsimonious model of such dependence is the matrix normal model, in which the covariances among the elements of a random matrix are parameterized in terms of the Kronecker product of two covariance matrices, one representing row covariances and one representing column covariance. An appealing feature of such a matrix normal model is that the Kronecker covariance structure allows for standard likelihood inference even when only a very small number of data matrices is available. For instance, in some cases a likelihood ratio test of dependence may be performed with a sample size of one. However, more generally the sample size required to ensure boundedness of the matrix normal likelihood or the existence of a unique maximizer depends in a complicated way on the matrix dimensions. This motivates the study of how large a sample size is needed to ensure that maximum likelihood estimators exist, and exist uniquely with probability one. Our main result gives precise sample size thresholds in the paradigm where the number of rows and the number of columns of the data matrices differ by at most a factor of two. Our proof uses invariance properties that allow us to consider data matrices in canonical form, as obtained from the Kronecker canonical form for matrix pencils.

Full work available at URL: https://arxiv.org/abs/2003.06024

Recommendations

• Existence and uniqueness of the maximum likelihood estimator for models with a Kronecker product covariance structure
• Gaussian and robust Kronecker product covariance estimation: existence and uniqueness
• Models with a Kronecker product covariance structure: estimation and testing
• Maximum likelihood estimation for matrix normal models via quiver representations
• Multivariate model with a Kronecker product covariance structure: S. N. Roy method of estimation

zbMATH Keywords

maximum likelihood estimationGaussian distributionKronecker canonical formseparable covariancematrix normal model

Mathematics Subject Classification ID

Estimation in multivariate analysis (62H12) Algebraic statistics (62R01)

Cites Work

• The mle algorithm for the matrix normal distribution
• Title not available (Why is that?)
• Title not available (Why is that?)
• Some matrix-variate distribution theory: Notational considerations and a Bayesian application
• Smooth nonlinear optimization of \(\mathbb R^n\)
• Gemini: graph estimation with matrix variate normal instances
• Separable factor analysis with applications to mortality data
• Geodesic Convexity and Covariance Estimation
• The multilinear normal distribution: introduction and some basic properties
• Transposable regularized covariance models with an application to missing data imputation
• Title not available (Why is that?)
• Testing for nodal dependence in relational data matrices
• Matrices and matroids for systems analysis
• Gaussian and robust Kronecker product covariance estimation: existence and uniqueness
• Model selection and estimation in the matrix normal graphical model
• A Geometric Approach to Perturbation Theory of Matrices and Matrix Pencils. Part I: Versal Deformations
• Are a set of microarrays independent of each other?
• Simplicity of core arrays in three-way principal component analysis and the typical rank of \(p\times q\times 2\) arrays
• Graph estimation for matrix-variate Gaussian data
• An expectation-maximization algorithm for the matrix normal distribution with an application in remote sensing
• Robust Kronecker Product PCA for Spatio-Temporal Covariance Estimation
• Inference with Transposable Data: Modelling the Effects of Row and Column Correlations
• The maximum likelihood threshold of a graph
• Multiple Matrix Gaussian Graphs Estimation
• The maximum likelihood threshold of a path diagram
• Invariant Theory and Scaling Algorithms for Maximum Likelihood Estimation
• Maximum likelihood estimation for matrix normal models via quiver representations

Cited In (9)

• Existence and uniqueness of the maximum likelihood estimator for models with a Kronecker product covariance structure
• Gaussian and robust Kronecker product covariance estimation: existence and uniqueness
• Maximum likelihood estimation for tensor normal models via castling transforms
• Maximum likelihood estimation for matrix normal models via quiver representations
• A bridge between invariant theory and maximum likelihood estimation
• Rational maximum likelihood estimators of Kronecker covariance matrices
• Data-driven optimal shrinkage of singular values under high-dimensional noise with separable covariance structure with application
• Fast and Separable Estimation in High-Dimensional Tensor Gaussian Graphical Models
• A note on necessary and sufficient conditions of existence and uniqueness for the maximum likelihood estimator of a Kronecker-product variance-covariance matrix

This page was built for publication: Existence and uniqueness of the Kronecker covariance MLE

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2054526)