A New Bayesian Network Structure for Classification Tasks

摘要

This paper introduces a new Bayesian network structure, named a Partial Bayesian Network (PBN), and describes an algorithm for constructing it. The PBN is designed to be used for classification tasks, and accordingly the algorithm constructs an approximate Markov blanket around a classification node. Initial experiments have compared the performance of the PBN algorithm with Naive Bayes, Tree-Augmented Naieve Bayes and a general Bayesian network algorithm (K2). The results indicate that PBN performs better than other Bayesian network classification structures on some problem domains.