A Heuristic Algorithm for Pattern-to-DAG Conversion

Yingying Wen and K.B. Korb (Australia)

Keywords

Algorithm, causal discovery algorithm, patterntoDAG conversion, evaluation

Abstract

Finding a DAG to represent a Markov equivalence class of DAGs — i.e., a pattern — is a necessary step in some causal discovery algorithms. If the case involves a known true DAG generating artificial data, then it is also arguably a necessary step in evaluating any causal discovery algo rithm. We present three algorithms for converting patterns to representative DAGs, analyse their time complexity and demonstrate their use experimentally.

Important Links:



Go Back