基于改进NB分类方法的网络异常检测模型

Network Anomaly Detection Model Based on Improved Nave Bayesian Classifier

随着计算机网络和分布式应用的复杂化和多样化,智能化网络异常检测技术逐渐成为有效监测和控制系统的重要方法。该文提出基于改进NB分类方法的网络异常检测算法,采用互信息的方法对网络属性进行关键特征提取。实验结果表明,该异常检测方法对DoS和Probing类攻击的检测率较高,具有较低的虚警率。

Due to the diversity and quickly growth of modem network and distributed systems, intelligent anomaly detection can play an important role in monitoring and controlling network systems. This paper presents a network anomaly detection model based on enhanced Nal＇ve Bayesian classifier. Mutual information in information theory is used as basis of the algorithm of feature selection. Experimental results show that the network anomaly detection model performs better in detection of DoS and probing attack, and has lower false positive.