Restricted Causal Inference Algorithm

TL;DR

This paper introduces a new algorithm for recovering belief network structures from data, especially handling hidden variables, by extending existing causal inference algorithms with restrictions and transformation steps.

Contribution

It extends the CI algorithm to handle hidden variables and transforms partial graphs into belief networks, with proven correctness.

Findings

Algorithm effectively recovers belief network structures.

Extensions improve handling of hidden variables.

Correctness of the method is demonstrated.

Abstract

This paper proposes a new algorithm for recovery of belief network structure from data handling hidden variables. It consists essentially in an extension of the CI algorithm of Spirtes et al. by restricting the number of conditional dependencies checked up to k variables and in an extension of the original CI by additional steps transforming so called partial including path graph into a belief network. Its correctness is demonstrated.