Jointly Interventional and Observational Data: Estimation of Interventional Markov Equivalence Classes of Directed Acyclic Graphs

DOI10.1111/RSSB.12071zbMATH Open1414.62021arXiv1303.3216OpenAlexW2138997118MaRDI QIDQ5379910

Alain Hauser, Peter Bühlmann

Publication date: 14 June 2019

Published in: Journal of the Royal Statistical Society Series B: Statistical Methodology (Search for Journal in Brave)

Abstract: In many applications we have both observational and (randomized) interventional data. We propose a Gaussian likelihood framework for joint modeling of such different data-types, based on global parameters consisting of a directed acyclic graph (DAG) and correponding edge weights and error variances. Thanks to the global nature of the parameters, maximum likelihood estimation is reasonable with only one or few data points per intervention. We prove consistency of the BIC criterion for estimating the interventional Markov equivalence class of DAGs which is smaller than the observational analogue due to increased partial identifiability from interventional data. Such an improvement in identifiability has immediate implications for tighter bounds for inferring causal effects. Besides methodology and theoretical derivations, we present empirical results from real and simulated data.

Full work available at URL: https://arxiv.org/abs/1303.3216

zbMATH Keywords

causal inferencemaximum likelihood estimationgraphical modelinterventionsBayesian information criteriongreedy equivalence search

Mathematics Subject Classification ID

Bayesian inference (62F15) Nonparametric estimation (62G05) Applications of statistics to biology and medical sciences; meta analysis (62P10) Estimation in multivariate analysis (62H12) Graphical methods in statistics (62A09)

Cited In (16)

• Compatible priors for model selection of high-dimensional Gaussian DAGs
• Objective Bayes model selection of Gaussian interventional essential graphs for the identification of signaling pathways
• Objective Bayesian Nets for Integrating Consistent Datasets
• Bayesian sample size determination for causal discovery
• Invariance, causality and robustness
• Learning Markov equivalence classes of directed acyclic graphs: an objective Bayes approach
• Equivalence class selection of categorical graphical models
• High-dimensional joint estimation of multiple directed Gaussian graphical models
• Causal deep learning: encouraging impact on real-world problems through causality
• Discussion to: Bayesian graphical models for modern biological applications by Y. Ni, V. Baladandayuthapani, M. Vannucci and F.C. Stingo
• Network modeling in biology: statistical methods for gene and brain networks
• Penalized Estimation of Directed Acyclic Graphs From Discrete Data
• Network Structure Learning Under Uncertain Interventions
• Inferring large graphs using \(\ell_1\)-penalized likelihood
• Beyond the mean: a flexible framework for studying causal effects using linear models
• Bayesian inference of causal effects from observational data in Gaussian graphical models

This page was built for publication: Jointly Interventional and Observational Data: Estimation of Interventional Markov Equivalence Classes of Directed Acyclic Graphs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5379910)