基于信息增益和粗糙集的入侵检测方法

传统的随机森林在网络入侵检测中收敛速度慢,并且学习性能不够完善。为消除原始入侵检测数据中的冗余信息,提出一种基于信息增益和粗糙集的随机森林入侵检测方法。使用信息增益对数据的各个属性进行相关分析,删除冗余属性,减小属性简约的时间复杂度;利用粗糙集理论从数据中提取分明函数,求得属性简约;使用随机森林分类器进行分类。实验结果表明,该方法收敛速度较快,在召回率和精度方面都要高于传统的随机森林方法,尤其是在训练样本充足的网络环境下,效果更加明显。

INTRUSION DETECTION METHOD BASED ON INFORMATION GAIN AND ROUGH SET

Traditional random forest converges slowly in network intrusion detection,and its learning performance is not perfect.In order to eliminate the redundant information in the original intrusion detection data,this paper proposes a random forest intrusion detection method based on information gain and rough set.It analyzed the correlation of data attributes with information gain,removed the redundant attributes,and reduced the time complexity of the attribute simplicity.The rough set theory was adopted to extract the distinct functions from the data to obtain the attribute simplicity.And we used a random forest classifier for classification.The experimental results show that the proposed method has a faster convergence rate and higher recall rate and accuracy than the traditional random forest method,especially in the network environment with sufficient training samples.