Distributional Equivalence and Structure Learning for Bow-free Acyclic\n Path Diagrams

2015/08/07 by Christopher Nowzohour, Nowzohour, Christopher, Marloes H. Maathuis +5 · 2 citations
Computer Science · #Advanced Database Systems and Queries #Bayesian Modeling and Causal Inference #FOS: Computer and information sciences #Machine Learning (stat.ML) #Statistical and Computational Modeling

paper · pdf · doi:10.48550/arxiv.1508.01717

openalex publication_date 2015/08/07 · openalex created_date 2022/09/17 · openalex updated_date 2026/07/28

Abstract

We consider the problem of structure learning for bow-free acyclic path\ndiagrams (BAPs). BAPs can be viewed as a generalization of linear Gaussian DAG\nmodels that allow for certain hidden variables. We present a first method for\nthis problem using a greedy score-based search algorithm. We also prove some\nnecessary and some sufficient conditions for distributional equivalence of BAPs\nwhich are used in an algorithmic ap- proach to compute (nearly) equivalent\nmodel structures. This allows us to infer lower bounds of causal effects. We\nalso present applications to real and simulated datasets using our publicly\navailable R-package.\n

Cited by

• Score-based Greedy Search for Structure Identification of Partially Observed Linear Causal Models
• Expert-aided causal discovery of ancestral graphs

Related

• Characterizing Distribution Equivalence and Structure Learning for\n Cyclic and Acyclic Directed Graphs
• Robust causal structure learning with some hidden variables
• Learning DAGs without imposing acyclicity
• ExDAG: an MIQP Algorithm for Learning DAGs
• Efficient Permutation Discovery in Causal DAGs