On Robust Trimming of Bayesian Network Classifiers

This paper considers the problem of removing costly features from a Bayesian network classifier. We want the classifier to be robust to these changes, and maintain its classification behavior. To this end, we propose a closeness metric between Bayesian classifiers, called the expected classification...

Full description

Saved in:
Bibliographic Details
Main Authors Choi, YooJung, Broeck, Guy Van den
Format Journal Article
LanguageEnglish
Published 29.05.2018
Subjects
Computer Science - Learning
Statistics - Machine Learning
Online AccessGet full text
DOI10.48550/arxiv.1805.11243

Cover

More Information
Summary:This paper considers the problem of removing costly features from a Bayesian network classifier. We want the classifier to be robust to these changes, and maintain its classification behavior. To this end, we propose a closeness metric between Bayesian classifiers, called the expected classification agreement (ECA). Our corresponding trimming algorithm finds an optimal subset of features and a new classification threshold that maximize the expected agreement, subject to a budgetary constraint. It utilizes new theoretical insights to perform branch-and-bound search in the space of feature sets, while computing bounds on the ECA. Our experiments investigate both the runtime cost of trimming and its effect on the robustness and accuracy of the final classifier.
DOI:10.48550/arxiv.1805.11243