基于神经网络的P2P流量识别方法  

P2P Traffic Identification Based on Neural Network

作　　者：李明[1] 贾波[1] 

机构地区：[1]浙江工商大学信息与电子工程学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2011年第4期152-156,共5页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：利用小波神经网络实时学习和快速识别的优点,该文提出一种统计特征和小波神经网络相结合的P2P流量识别方法。在实际网络环境下,通过建立网络分类模型,统计分析并提取多种流量特性,通过小波神经网络对各种P2P应用流量特征的学习和识别,提高了P2P流量识别的准确度,改善了之前单一识别方法的复杂度。Real-time using neural network to quickly identify the advantages of learning and presents a wave- let neural network and statlstical characteristics of the combination of P2P traffic identification method. In real network environment, through the establishment of the network classification model, statistical analysis and ex- tract a variety of flow characteristics, the wavelet neural network P2P application traffic characteristics of vari- ous learning and recognition, improve the accuracy of P2P traffic identification, improving the previous one Reeognition of the complexity.

关 键 词：统计学习 神经网络 流量识别 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]