Gaussoids are two-antecedental approximations of Gaussian conditional independence structures

DOI10.1007/S10472-021-09780-0zbMATH Open1493.62020arXiv2010.11914OpenAlexW3215476779WikidataQ114227640 ScholiaQ114227640MaRDI QIDQ2149807FDOQ2149807

Authors: Tobias Boege 

Publication date: 29 June 2022

Published in: Annals of Mathematics and Artificial Intelligence (Search for Journal in Brave)

Abstract: The gaussoid axioms are conditional independence inference rules which characterize regular Gaussian CI structures over a three-element ground set. It is known that no finite set of inference rules completely describes regular Gaussian CI as the ground set grows. In this article we show that the gaussoid axioms logically imply every inference rule of at most two antecedents which is valid for regular Gaussians over any ground set. The proof is accomplished by exhibiting for each inclusion-minimal gaussoid extension of at most two CI statements a regular Gaussian realization. Moreover we prove that all those gaussoids have rational positive-definite realizations inside every $varepsilon$-ball around the identity matrix. For the proof we introduce the concept of algebraic Gaussians over arbitrary fields and of positive Gaussians over ordered fields and obtain the same two-antecedental completeness of the gaussoid axioms for algebraic and positive Gaussians over all fields of characteristic zero as a byproduct.

Full work available at URL: https://arxiv.org/abs/2010.11914

Recommendations

• The geometry of gaussoids
• On the completeness of the semigraphoid axioms for deriving arbitrary from saturated conditional independence statements
• Gaussian conditional independence relations have no finite complete characterization
• On Gaussian conditional independence structures
• Conditionals, Information, and Inference

zbMATH Keywords

conditional independenceinferencerationalitycompletenessrealizabilitygaussoid

Mathematics Subject Classification ID

Statistical aspects of information-theoretic topics (62B10) Algebraic statistics (62R01) Semialgebraic sets and related spaces (14P10)

Cites Work

• Ideals, Varieties, and Algorithms
• Title not available (Why is that?)
• On the conditional independence implication problem: a lattice-theoretic approach
• Efficient algorithms for conditional independence inference
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Logical and algorithmic properties of conditional independence and graphical models
• Positive polynomials and sums of squares
• On Gaussian conditional independence structures
• Hyperdeterminantal relations among symmetric principal minors
• Conditional Independences among Four Random Variables I
• Towards classification of semigraphoids.
• Conditional Independences among Four Random Variables III: Final Conclusion
• Conditionals, Information, and Inference
• Smoothness of Gaussian conditional independence models
• Algebraic statistics
• Gaussian conditional independence relations have no finite complete characterization
• PROBABILISTIC CONDITIONAL INDEPENDENCE STRUCTURES AND MATROID THEORY: BACKGROUND¹
• Conditional independence structures examined via minors
• Conditional Independences among Four Random Variables II
• Unifying Markov properties for graphical models
• The geometry of gaussoids
• Conditional independence and basic Markov properties
• Construction methods for gaussoids.
• A note on a generalized Cramer's rule

Cited In (4)

• The geometry of Gaussian double Markovian distributions
• Construction methods for gaussoids.
• Self-adhesivity in lattices of abstract conditional independence models
• From matrix polynomial to determinant of block Toeplitz-Hessenberg matrix

Uses Software

• Macaulay2

This page was built for publication: Gaussoids are two-antecedental approximations of Gaussian conditional independence structures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2149807)