Alert button

On Robust Trimming of Bayesian Network Classifiers

Add code
Alert button
May 29, 2018
Figure 1 for On Robust Trimming of Bayesian Network Classifiers
Figure 2 for On Robust Trimming of Bayesian Network Classifiers
Figure 3 for On Robust Trimming of Bayesian Network Classifiers
Figure 4 for On Robust Trimming of Bayesian Network Classifiers

Share this with someone who'll enjoy it:

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.

Share this with someone who'll enjoy it: