Maximum likelihood estimation in Gaussian models under total positivity
From MaRDI portal
Abstract: We analyze the problem of maximum likelihood estimation for Gaussian distributions that are multivariate totally positive of order two (MTP2). By exploiting connections to phylogenetics and single-linkage clustering, we give a simple proof that the maximum likelihood estimator (MLE) for such distributions exists based on at least 2 observations, irrespective of the underlying dimension. Slawski and Hein, who first proved this result, also provided empirical evidence showing that the MTP2 constraint serves as an implicit regularizer and leads to sparsity in the estimated inverse covariance matrix, determining what we name the ML graph. We show that we can find an upper bound for the ML graph by adding edges corresponding to correlations in excess of those explained by the maximum weight spanning forest of the correlation matrix. Moreover, we provide globally convergent coordinate descent algorithms for calculating the MLE under the MTP2 constraint which are structurally similar to iterative proportional scaling. We conclude the paper with a discussion of signed MTP2 distributions.
Recommendations
- Maximum likelihood estimation for linear Gaussian covariance models
- scientific article; zbMATH DE number 775123
- Maximum likelihood estimation for Gaussian processes under inequality constraints
- Maximum Likelihood Estimation for a Smooth Gaussian Random Field Model
- Maximum likelihood estimation in linear infinite dimensional models
- Positive definite maximum likelihood covariance estimators
- scientific article; zbMATH DE number 3886910
Cites work
- scientific article; zbMATH DE number 6484043 (Why is no real title available?)
- scientific article; zbMATH DE number 5957391 (Why is no real title available?)
- scientific article; zbMATH DE number 3770781 (Why is no real title available?)
- scientific article; zbMATH DE number 469396 (Why is no real title available?)
- scientific article; zbMATH DE number 1134987 (Why is no real title available?)
- scientific article; zbMATH DE number 1865935 (Why is no real title available?)
- A general central limit theorem for FKG systems
- A likelihood ratio test for \(\text{MTP}_ 2\) within binary variables.
- A recursive algorithm for Markov random fields
- Classes of orderings of measures and related correlation inequalities. I. Multivariate totally positive distributions
- Correlation inequalities on some partially ordered sets
- Covariance selection and multivariate dependence
- Estimation of positive definite \(M\)-matrices and structure learning for attractive Gaussian Markov random fields
- Exact sampling with coupled Markov chains and applications to statistical mechanics
- Gaussian Markov distributions over finite graphs
- Geometry of maximum likelihood estimation in Gaussian graphical models
- Graphical models with R.
- High-dimensional Gaussian graphical model selection: walk summability and local separation criterion
- Inverse \(M\)-matrices and ultrametric matrices
- Inverse \(M\)-matrices. II
- Learning Graphs With Monotone Topology Properties and Multiple Connected Components
- Learning latent tree graphical models
- M-matrices as covariance matrices of multinormal distributions
- On the convergence of the coordinate descent method for convex differentiable minimization
- Path product matrices
- Some notions of multivariate positive dependence
- The completion problem for \(M\)-matrices and inverse \(M\)-matrices
- The correlation space of Gaussian latent tree models and model selection without fitting
- The maximum likelihood threshold of a graph
- Total positivity in Markov structures
- Total positivity properties of absolute value multinormal variables with applications to confidence interval estimates and related probabilistic inequalities
- Towards a Unified Theory of Inequality Constrained Testing in Multivariate Analysis
- Über die Determinanten mit überwiegender Hauptdiagonale
Cited in
(22)- Total positivity in Markov structures
- Convolutions of totally positive distributions with applications to kernel density estimation
- Gaussian Markov random fields and totally positive matrices
- Total positivity in exponential families with application to binary variables
- On Sidorenko's conjecture for determinants and Gaussian Markov random fields
- Tropical Gaussians: a brief survey
- The geometry of gaussoids
- Estimation of Monge matrices
- Likelihood-based estimation for Gaussian MRFs
- Nonlinear algebra and applications
- A multitest procedure for testing MTP2 for Gaussian distributions
- Universal inference
- Locally associated graphical models and mixed convex exponential families
- Maximum likelihood estimation for totally positive log-concave densities
- Maximum likelihood degree of the two-dimensional linear Gaussian covariance model
- Optimal rates for estimation of two-dimensional totally positive distributions
- Bimonotone subdivisions of point configurations in the plane
- Learning graph Laplacian with MCP
- Maximum likelihood geometry in the presence of data zeros
- Total positivity in multivariate extremes
- High-Order Analysis of the Efficiency Gap for Maximum Likelihood Estimation in Nonlinear Gaussian Models
- Dependence in elliptical partial correlation graphs
This page was built for publication: Maximum likelihood estimation in Gaussian models under total positivity
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2313271)
