摘要
P2P技术作为一种全新的网络应用,正主导着互联网的发展方向,P2P的管理问题也成为当前互联网络中最大的难题。通过分析P2P流量特征及控制P2P流量过程中存在的问题,比较目前P2P流量检测的几种技术,提出一种基于属性关键度的多决策树分类方法,设计了一个基于多决策树算法的P2P流量检测模型,阐述了模型的工作原理。从虚警率和漏警率以及检测率三个方面评价了采用多决策树算法进行P2P流量检测的有效性。通过大量实验证明,该方法具有较高的检测率,说明采用多决策树分类算法进行P2P流量检测的有效性。
As a new network application,P2P technology is leading the direction of development of the Internet.The management of P2P network has also become the biggest problem in the Internet.By analyzing the P2P traffic features and the problems in the process of controlling P2P traffic and comparing existing technologies about P2P traffic detection,proposed a decision-tree algorithm based on the degree of important of property,designed a P2P detection model based on multi-decision tree algorithm,and also described the working principle of model. Evaluate the classification in term of false positive,false negative and detection rate.The experiment showed that the method has a high identification rate to the protocol of P2P,which showed that an effectiveness in using decision -tree classification algorithm on the identification of P2P traffic.
出处
《计算机技术与发展》
2010年第6期126-129,共4页
Computer Technology and Development
关键词
对等网络
流量检测
数据挖掘
决策树算法
peer to peer
traffic detection
data mining
decision tree algorithm
