Margins of discrete Bayesian networks
From MaRDI portal
Publication:1991671
Abstract: Bayesian network models with latent variables are widely used in statistics and machine learning. In this paper we provide a complete algebraic characterization of Bayesian network models with latent variables when the observed variables are discrete and no assumption is made about the state-space of the latent variables. We show that it is algebraically equivalent to the so-called nested Markov model, meaning that the two are the same up to inequality constraints on the joint probabilities. In particular these two models have the same dimension. The nested Markov model is therefore the best possible description of the latent variable model that avoids consideration of inequalities, which are extremely complicated in general. A consequence of this is that the constraint finding algorithm of Tian and Pearl (UAI 2002, pp519-527) is complete for finding equality constraints. Latent variable models suffer from difficulties of unidentifiable parameters and non-regular asymptotics; in contrast the nested Markov model is fully identifiable, represents a curved exponential family of known dimension, and can easily be fitted using an explicit parameterization.
Recommendations
- Smooth, identifiable supermodels of discrete DAG models with latent variables
- Graphs for margins of Bayesian networks
- Identification of discrete concentration graph models with one hidden binary variable
- Algebraic geometry of Bayesian networks
- Geometry, moments and conditional independence trees with hidden variables.
Cites work
- scientific article; zbMATH DE number 1055955 (Why is no real title available?)
- A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect
- Algebraic geometry of Bayesian networks
- Algorithms in real algebraic geometry
- Ancestral graph Markov models.
- Causality. Models, reasoning, and inference
- Causation, prediction, and search
- Decompositions of Simplicial Complexes Related to Diameters of Convex Polyhedra
- Graphs for margins of Bayesian networks
- Ideals, varieties, and algorithms. An introduction to computational algebraic geometry and commutative algebra
- Identifiability of parameters in latent structure models with many observed variables
- Influence Diagrams for Causal Modelling and Inference
- Likelihood ratio tests and singularities
- Markov Properties for Acyclic Directed Mixed Graphs
- Markovian acyclic directed mixed graphs for discrete data
- Modeling and Reasoning with Bayesian Networks
- On the causal interpretation of acyclic mixed graphs under multivariate normality
- Pattern recognition and machine learning.
- Statistics, causality and Bell's theorem
- The hidden life of latent variables: Bayesian learning with mixed graph models
- Theory-independent limits on correlations from generalized Bayesian networks
Cited in
(13)- Nested Markov properties for acyclic directed mixed graphs
- Decision boundary for discrete Bayesian network classifiers
- A study of recently discovered equalities about latent tree models using inverse edges
- Smooth, identifiable supermodels of discrete DAG models with latent variables
- Discretisation Does Affect the Performance of Bayesian Networks
- Graphs for margins of Bayesian networks
- Latent free equivalent mDAGs
- An Automated Approach to Causal Inference in Discrete Settings
- Marginal Markov subgraph of Bayesian network and its applications
- Foundations of structural causal models with cycles and latent variables
- Analysis of Markov boundary induction in Bayesian networks: a new view from matroid theory
- Causal structure learning: a combinatorial perspective
- Discrete max-linear Bayesian networks
This page was built for publication: Margins of discrete Bayesian networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1991671)
