A Network Traffic Classification Model Based On XGBOOST_RFECV Feature Extraction

Abstract

Network traffic plays a crucial role in the interaction and transfer of information in the network area, which contains a large amount of information with important value. Therefore, network traffic classification is essential for network management, security monitoring and intrusion detection. However, the performance of network traffic classification is greatly affected by the extremely unbalanced datasets which are publicly available. In order to solve the problem of low accuracy of minority class classification. In this paper, we used SMOTEENN as the balanced method and XGBOOST_RFECV was used for feature selection. Subsequently, the neural network model (1DCNN_BiLSTM) was used for training and verification. The experimental results show that this method can effectively solve the problem of imbalanced data category, which has certain reference significance for the research of network traffic classification technology.