Abstract: This paper discusses Naive Bayes classification theory and methods,which can be applied to textual data analyzing.In this paper we analyze network security audit data with Naive Bayesian classifier.In order to improve the efficiency of classification,attributes that lead to better classification result are selected,and the other are aborted.Tests on various attributes of KDD CUP99 data sets show that classification results are rather good.

摘要： 研究了非监督学习Na?觙ve Bayes分类的原理和方法,并将其应用到文本数据——网络安全审计数据的分析中。为了提高分类准确率,根据分类的效果对数据的属性集进行选择,使用能提高分类准确性的属性作为分类的依据。对KDD CUP99数据集进行了基于不同属性集的实验,发现了与分类结果相关的属性,分类效果良好。