摘要： 介绍增量贝叶斯分类器的原理，提出一种基于类支持度的优化增量贝叶斯分类器学习算法。在增量学习过程的样本选择问题上，算法引入一个类支持度因子λ，根据λ的大小逐次从测试样本集中选择样本加入分类器。实验表明，在训练数据集较小的情况下，该算法比原增量贝叶斯分类算法具有更高的精度，能大幅度减少增量学习样本优选的计算时间。

关键词: 贝叶斯分类器, 分类算法, 增量学习

Abstract: A learning incremental Bayesian classifier algorithm based on class support is presented. In the sample selection process of the incremental learning algorithm, a kind of support factor λ is introduced. According the size of λ, the samples are selected from test sample aggregation and are joined in the classifier gradually. The experiments show that this algorithm has higher precision compared with simple incremental Bayesian classifier with small training data set and it can reduce largely the computing time that costs in samples optimal selection in incremental learning.

Key words: Bayesian classifier, classification algorithm, incremental learning

中图分类号: 

• TP301