Research on Malicious Traffic Detection Technology Based on Deep Learning

For network traffic, how to distinguish malicious traffic from normal traffic is very important. To this end, we introduce deep learning as a powerful tool for malicious traffic detection, and its excellent learning ability shines in this field. This paper proposes a double-layer neural network model based on deep learning, which combines DCNN and LSTM networks to detect and classify network traffic. The UNSW_NB15 dataset was used, LabelEncoder function and MAX-MIN standardization were used to preprocess the data, ADASYN was used to balance the data, random forest algorithm and Pearson correlation analysis were used to select the features, and 27 features were finally selected for classification. DCNN extracts features, LSTM builds a sequence model to deal with long-term dependencies, and finally, softmax classifier is used. Compared with the single DCNN and LSTM structure, the double-layer neural network model shows a good classification effect.