With the development of information technology,the peer-to-peer(P2P)network information traffic often deviates from the normal range. The detection technology for P2P traffic detection and abnormal traffic is studied on the basis of the deci-sion tree algorithm. The P2P traffic detection model based on improved C4.5 decision tree is used to train the massive training datasets by means of the P2P anomaly traffic detection model to modify the error gradually. The simulation test in laboratory was performed. The P2P network traffic classifier based on improved C4.5 decision tree has perfect classification effect after selecting the characteristics of the network traffic. The classification detection rate is 94.6%~96.7%,which shows that the improved C4.5 decision tree algorithm can detect the P2P traffic effectively,and provide the reference for studying the P2P anomaly traffic detection technology in future.%随着信息技术的发展,对等网络P2P信息流量经常出现偏离正常范围的异常情况,这里以决策树算法为基础,对P2P流量检测和流量异常时的检测技术进行研究.采用改进的C4.5决策树P2P流量检测模型,通过P2P流量异常检测模型对大量训练数据集的训练,实现了对错误的逐步修正,通过试验室仿真试验可知,经过选择网络流量特征后,基于改进的C4.5决策树的P2P网络流量分类器能实现较好的分类效果,分类检测率在94.6%~96.7%,较高的检测率说明采用改进的C4.5决策树算法能有效地对P2P流量进行检测,为研究P2P流量异常检测技术提供了参考.
展开▼
