An Abnormal Network Traffic Detection Method on MAWILab Dataset based on Convolutional Neural Network
Xin Liu · 2022 IEEE 2nd International Conference on Electronic Technology, Communication and Information (ICETCI) · 2022
In today's complex network environment, there are many data sources from various network nodes, and it is very difficult to detect network security events. In order to quickly and effectively defend against network attacks, effective technical means are needed to aggregate network threat data, and to implement detection and defense as soon as possible. In recent years, network threat intelligence detection based on machine learning has been widely studied and used, which can provide better defense against new attacks with higher accuracy and higher detection rate. This paper studies and analyzes the network traffic data in the traffic database MAWILab data set, filters out the bidirectional traffic of the sessions containing all layers, splits the traffic data and classifies the labels, divides the training sets and the test sets, and then trains and tests traffic data using convolutional neural networks. The results show that in the MAWILab dataset, the detection effect of the convolutional neural network based on session bidirectional traffic is ideal, and it can achieve a satisfactory detection effect in the case of the limited resources.