摘要
朴素贝叶斯分类器(Naive Bayesian classmer,NB)是一种简单而有效的分类模型,但这种分类器缺乏对训练集信息的充分利用,影响了它的分类性能。通过分析NB的分类原理,并结合线性判别分析(Linear Discriminant A- nalysis,LDA)与核判别分析(Kemel Discriminant Analysis,KDA)的优点,提出了一种混合贝叶斯分类模型DANB (Discriminant Analysis Naive Bayesian classifier,DANB)。将该分类方法与NB和TAN(Tree Augmented Naive Bayesian classifier,TAN)进行实验比较,结果表明,在大多数数据集上,DANB分类器具有较高的分类正确率。
Naive Bayesian classifier (NB) is a simple and effective classification model, but it is unable to make the best of the information of the training dataset, thus affecting its classification performance. On the basis of analyzing the classification principle of NB and integrating strongpoint of Linear Discriminant Analysis (LDA) and Kernel Discriminant Analysis (KDA), a new hybrid Bayesian classification model, DANB (Discriminant Analysis Naive Bayesian classifter), is proposed. DANB classifier is compared with NB and TAN (Tree Augmented Naive Bayesian classifier) by an experiment. Experiment results show that this model has higher classification accuracy in most datasets.
出处
《计算机科学》
CSCD
北大核心
2006年第9期135-139,共5页
Computer Science
基金
国家自然科学基金资助课题(70371026)。
关键词
朴素贝叶斯分类器
线性判别分析
核判别分析
TAN分类器
Naive Bayesian classifier,Linear discriminant analysis,Kernel discriminant analysis,TAN classification