ExpertBayes: Automatically refining manually built Bayesian networks

TL;DR

ExpertBayes automatically refines expert-designed Bayesian networks through minor perturbations, improving classifier performance efficiently while preserving the original domain knowledge, especially useful in fields like medicine.

Contribution

This work introduces a method to automatically refine manually built Bayesian networks with minimal computational effort, enhancing their predictive accuracy.

Findings

Minor perturbations improve classifier performance

Refinement maintains original domain knowledge

Method is computationally efficient

Abstract

Bayesian network structures are usually built using only the data and starting from an empty network or from a naive Bayes structure. Very often, in some domains, like medicine, a prior structure knowledge is already known. This structure can be automatically or manually refined in search for better performance models. In this work, we take Bayesian networks built by specialists and show that minor perturbations to this original network can yield better classifiers with a very small computational cost, while maintaining most of the intended meaning of the original model.