Malicious network traffic detection method based on traffic behavior characteristics and machine learning

This paper first analyzes the research background and current situation of unknown network threat detection, points out the main problems existing in the current method, and clarifies the research scope and basic ideas of this paper. On this basis, the main research contents and innovations of this paper are as follows: 1) In response to the lack of implementation ideas for detecting unknown network threats in large-scale network systems, a coarse-grained extraction approach for network traffic load characteristics is proposed. Based on the load, it provides the most valuable feature information for subsequent machine learning models, reduces learning costs, and improves model fitting ability; 2) For coarse-grained load information, we use fine-grained text2vec statistical feature construction techniques for processing; 3) Based on the text2vec feature vectors, we select the optimal dimension; 4) Using machine learning models based on high-dimensional statistical features to perform complex calculations and output classification results.