海量冗余数据冲击下网络入侵检测方法

Network intrusion detection method under the impact of massive redundant data

针对冗余数据冲击下网络入侵数量较大,无法很好地区分正常行为和入侵行为,导致检测准确性差的问题,该文提出一种海量冗余数据冲击下网络入侵检测方法。利用独热编码方式对网络入侵数据进行编码处理,将其转换为数值类型特征。采用PCA算法对海量冗余数据进行压缩,通过数据聚类算法实现网络入侵压缩数据采样,改善海量冗余数据质量,去除网络数据中的噪声;通过改进K-means算法划分数据类簇,筛选内部异常点;利用支持向量机方法为每个类簇构建单分类器,进而判别攻击类型,减少误报率,通过梯度下降法实现网络入侵数据的寻优,实现海量冗余数据冲击下的网络入侵检测。经实验测试结果表明,所提方法可以有效降低网络入侵检测时间,提升检测结果的准确性。

In view of the large number of network intrusion under the impact of redundant data,which can not well distinguish between normal behavior and intrusion behavior,resulting in poor detection accuracy,this paper proposes a network intrusion detection method under the impact of massive redundant data.The unique hot coding method is used to encode the network intrusion data and convert it into numerical type features.PCA algorithm is used to compress massive redundant data,and data clustering algorithm is used to realize network intrusion compressed data sampling,improve the quality of massive redundant data and remove the noise in network data.The improved K-means algorithm is used to divide data clusters and screen internal outliers.The support vector machine method is used to construct a single classifier for each class cluster,so as to distinguish the attack type and reduce the false positive rate.The gradient descent method is used to optimize the network intrusion data and realize the network intrusion detection under the impact of massive redundant data.The experimental results show that the proposed method can effectively reduce the network intrusion detection time and improve the accuracy of the detection results.