Boosting Local Naïve Bayesian Rules

  • Authors:
  • Zhipeng Xie

  • Affiliations:
  • School of Computer Science, Fudan University, Shanghai, China 200433

  • Venue:
  • ISNN 2009 Proceedings of the 6th International Symposium on Neural Networks: Advances in Neural Networks - Part II
  • Year:
  • 2009

Quantified Score

Hi-index 0.01

Visualization

Abstract

Several classification algorithms based on local naïve Bayesian rules have been recently developed to provide high predictability. However, most of them use classifier selection strategy in decision making. To make use of classifier fusion strategy, this paper investigates a boosting algorithm for local naïve Bayesian rules. Firstly, we develop an algorithmic framework as a forward stage-wise additive model. Then, a construction algorithm for lazy naïve Bayesian rules is designed to materialize the algorithmic framework. The construction algorithm starts from the most general rule, and uses a greedy search to grow the antecedent repeatedly in order to get a better rule at each step. Experimental results show that the proposed method has successfully reduced the overall error rate on a variety of domains, compared with boosted naïve Bayesian classifier, and lazy Bayesian rule algorithm.