基于CART的多属性度量报文分类算法

A CART-Based Packet Classification Algorithm for Multi-attribute Measurement

对于网络入侵检测、负载平衡、拥塞控制、QoS等各种网络应用技术,虽然实现的具体细节不同,但一个公共的要求是路由器能够基于报文头的某些字段对报文进行分类.提出了一种基于CART决策树的报文分类算法,采用了信息增益、增益率和Gini 3个指标综合考虑求解属性选择度量,与传统的报文分类匹配算法相比较,在精确性和匹配效率上都有较大提高.

For various network application technologies such as network-based intrusion detection systems,loading balance,congestion control and QoS,they all need packet classification based on some fields of packet header,although implementation of these functions varies greatly.A packet classification algorithm based on CART（classification and regression tree） is put forward in the paper.As it includes three data（information gain,information gain ratio and Gini） to solve attribute selection measurement,this packet classification algorithm improves both accuracy and matching efficiency considerably,compared with some traditional packet classification matching algorithms.