Identifiability of directed Gaussian graphical models with one latent source
From MaRDI portal
Abstract: We study parameter identifiability of directed Gaussian graphical models with one latent variable. In the scenario we consider, the latent variable is a confounder that forms a source node of the graph and is a parent to all other nodes, which correspond to the observed variables. We give a graphical condition that is sufficient for the Jacobian matrix of the parametrization map to be full rank, which entails that the parametrization is generically finite-to-one, a fact that is sometimes also referred to as local identifiability. We also derive a graphical condition that is necessary for such identifiability. Finally, we give a condition under which generic parameter identifiability can be determined from identifiability of a model associated with a subgraph. The power of these criteria is assessed via an exhaustive algebraic computational study on models with 4, 5, and 6 observable variables.
Recommendations
- Parameter identifiability in a class of random graph mixture models
- A Scalable Strategy for the Identification of Latent-Variable Graphical Models
- Identification of discrete concentration graph models with one hidden binary variable
- Directed Gaussian graphical models with toric vanishing ideals
- Maximum Likelihood Estimation Over Directed Acyclic Gaussian Graphs
- Identification in graphical causal models
- Prior distributions for Gaussian graphical models: a comparison between the directed and undirected case
- AR Identification of Latent-Variable Graphical Models
- High-dimensional joint estimation of multiple directed Gaussian graphical models
Cites work
- Algebraic factor analysis: tetrads, pentads and beyond
- Algorithms in real algebraic geometry
- Ancestral graph Markov models.
- Causality. Models, reasoning, and inference
- Generic identifiability of linear structural equation models by ancestor decomposition
- Global identifiability of linear structural equation models
- Graphical Identifiability Criteria for Total Effects in Studies with an Unobserved Response Variable
- Gröbner bases and triangulations of the second hypersimplex
- Half-trek criterion for generic identifiability of linear structural equation models
- scientific article; zbMATH DE number 45532 (Why is no real title available?)
- scientific article; zbMATH DE number 3539473 (Why is no real title available?)
- scientific article; zbMATH DE number 3573250 (Why is no real title available?)
- Lectures on algebraic statistics
- Linear Statistical Inference and its Applications
- Measurement bias and effect restoration in causal inference
- Miscellanea. Identification of a single-factor model using graphical Gaussian rules
- Miscellanea. On the identification of a single-factor model with correlated residuals
- On identification of multi-factor models with correlated residuals
- On the identification of path analysis models with one hidden variable
- Stratified exponential families: Graphical models and model selection
- The rank of reduced dispersion matrices
Cited in
(8)- Convergence rates of latent topic models under relaxed identifiability conditions
- Identifiability of phylogenetic parameters from k-mer data under the coalescent
- Smooth, identifiable supermodels of discrete DAG models with latent variables
- Spanning trees and identifiability of a single-factor model
- Half-trek criterion for identifiability of latent variable models
- Sparse directed acyclic graphs incorporating the covariates
- Identification of discrete concentration graph models with one hidden binary variable
- Miscellanea. Identification of a single-factor model using graphical Gaussian rules
This page was built for publication: Identifiability of directed Gaussian graphical models with one latent source
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5965329)
