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.
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
关键词
统计学习
神经网络
流量识别
statistical learning
neural network
traffic identification
