Scoring and Searching over Bayesian Networks with Causal and Associative Priors

TL;DR

This paper introduces a method for incorporating causal and associative prior beliefs into Bayesian network learning, improving the accuracy of network structure discovery through a novel search operator and prior assignment based on path beliefs.

Contribution

It presents a new approach for assigning priors based on path beliefs and introduces a novel search operator to leverage this prior knowledge in Bayesian network learning.

Findings

Path beliefs improve skeleton learning accuracy

Prior knowledge enhances edge direction determination

Method outperforms traditional approaches in experiments

Abstract

A significant theoretical advantage of search-and-score methods for learning Bayesian Networks is that they can accept informative prior beliefs for each possible network, thus complementing the data. In this paper, a method is presented for assigning priors based on beliefs on the presence or absence of certain paths in the true network. Such beliefs correspond to knowledge about the possible causal and associative relations between pairs of variables. This type of knowledge naturally arises from prior experimental and observational data, among others. In addition, a novel search-operator is proposed to take advantage of such prior knowledge. Experiments show that, using path beliefs improves the learning of the skeleton, as well as the edge directions in the network.