摘要
提出了基于DBSCAN算法的网络流量分类方法,对流的定义、特征产生、特征选择以及分类规则和分类性能的评测等内容进行了介绍。提出了基于PCA的网络流量最优特征子集的选择方法。实验结果表明,提出的分类方法能够达到较高的总精确度和查准率,能够有效地使用于网络流量分类中。
This paper presented a network traffic classification method based on DBSCAN algorithm, and introduced the definition of flow, the feature generated, the feature selection as well as the rule of classification and the performance evaluation of classification. Furthermore, employed the principle component analysis （PCA） approach to extract the optimization attribute set from the original network traffic data. The experiment results show that the method of presented can achieve higher overall accuracy and precision, and to effective use in network traffic classification.
出处
《计算机应用研究》
CSCD
北大核心
2009年第9期3461-3464,共4页
Application Research of Computers
基金
中国博士后科学基金资助项目(20070410299)
广东省自然科学基金博士科研启动基金资助项目(7300450)
关键词
网络流量分类
主成分分析
特征选择
DBSCAN聚类
network traffic classification
principle component analysis
feature selection
DBSCAN clustering
