Guided structure learning of DAGs for count data

TL;DR

This paper introduces a new algorithm for learning DAG structures from count data by leveraging known variable orderings, with proven consistency and competitive performance in finite samples.

Contribution

It proposes a novel DAG structure learning algorithm that incorporates prior topological order knowledge and proves its theoretical consistency.

Findings

Algorithm is consistent as sample size approaches infinity.

Experimental results show competitive performance with existing methods.

The approach effectively utilizes prior knowledge to improve structure learning.

Abstract

In this paper, we tackle structure learning of Directed Acyclic Graphs (DAGs), with the idea of exploiting available prior knowledge of the domain at hand to guide the search of the best structure. In particular, we assume to know the topological ordering of variables in addition to the given data. We study a new algorithm for learning the structure of DAGs, proving its theoretical consistence in the limit of infinite observations. Furthermore, we experimentally compare the proposed algorithm to a number of popular competitors, in order to study its behavior in finite samples.