Smoothness of Gaussian conditional independence models

MaRDI QIDQ3581603zbMATH OpenFDO

Authors Mathias Drton, Han Xiao

Publication date 2 September 2010

Full work available at URL https://arxiv.org/abs/0910.5447

zbMATH Keywords

multivariate normal distributionalgebraic statisticssingularities

Mathematics Subject Classification ID

Applications of graph theory (05C90) Characterization and structure theory for multivariate probability distributions; copulas (62H05)

Abstract: Conditional independence in a multivariate normal (or Gaussian) distribution is characterized by the vanishing of subdeterminants of the distribution's covariance matrix. Gaussian conditional independence models thus correspond to algebraic subsets of the cone of positive definite matrices. For statistical inference in such models it is important to know whether or not the model contains singularities. We study this issue in models involving up to four random variables. In particular, we give examples of conditional independence relations which, despite being probabilistically representable, yield models that non-trivially decompose into a finite union of several smooth submodels.

Recommendations

• Smoothness of conditional independence models for discrete data
• On Gaussian conditional independence structures
• Conditional independence in a Gaussian distribution.
• Conditionals, Information, and Inference
• Gaussian conditional independence relations have no finite complete characterization

Cited in

(15)

• Positivity for Gaussian graphical models
• On Gaussian conditional independence structures
• Conjunction probability of smooth centered Gaussian processes
• Smoothness of conditional independence models for discrete data
• Generalized Permutohedra from Probabilistic Graphical Models
• The geometry of gaussoids
• A class of smooth models satisfying marginal and context specific conditional independencies
• The geometry of Gaussian double Markovian distributions
• Conditional independence in a Gaussian distribution.
• On finite exchangeability and conditional independence
• On patterns of conditional independences and covariance signs among binary variables
• Markovian acyclic directed mixed graphs for discrete data
• On conditional independence and log-convexity
• Gaussoids are two-antecedental approximations of Gaussian conditional independence structures
• Smoothness of marginal log-linear parameterizations

Describes a project that uses

Uses Software

• SINGULAR

This page was built for publication: Smoothness of Gaussian conditional independence models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3581603)