基于支持向量机的通信网络攻击行为分类检测方法

Classification and Detection Method of Communication Network Attack Behavior Based on Support Vector Machine

由于通信网络流量数据具有高维性与复杂性,传统的网络攻击行为检测方法的检测准确率较低。为提高检测准确率,文章提出基于支持向量机的通信网络攻击行为分类检测方法,即使用预处理后的流量数据构建图卷积神经网络模型,提取特征并输入支持向量机进行分类,得到最终的攻击行为分类结果。仿真实验结果表明,基于图卷积神经网络的分类检测结果的漏报率仅为0.78%,相较于基于BP神经网络和基于普通卷积神经网络的分类检测方法具有更高的检测精度。

Due to the high dimensionality and complexity of communication network traffic data,traditional methods for detecting network attack behavior have lower accuracy.In order to improve the detection accuracy,this paper proposes a classification and detection method of communication network attack behavior based on support vector machines.This method uses preprocessed traffic data to construct a graph convolutional neural network model,extracts features,and inputs them into support vector machines for classification,obtaining the final classification result of attack behavior.The simulation experiments results show that the false positive rate of classification detection results based on graph convolutional neural networks is only 0.78%,which has higher detection accuracy compared to the classification detection methods based on BP neural networks and ordinary convolutional neural networks.