Network Security Entity Recognition Methods Based on the Deep Neural Network

Abstract

The network security threat intelligence analysis based on the security knowledge atlas can analyze the multi-source threat intelligence data in the fine granularity, which attracts the wide attention. The traditional named entity recognition methods are difficult to identify the new or the mixed Chinese and English security entities in the field of the network security, and the extracted features are inadequate, so it is difficult to accurately identify the network security entities. On the basis of the deep neural network model, a method of the network security entity recognition based on CNN-BiLSTM-CRF combined with the feature template is proposed. The artificial feature template is used to extract the local context features, and then the neural network model is used to automatically extract the character features and the text global features. The experimental results show that on the large-scale network security data sets, the proposed method of the network security entity recognition is effective. The relevant evaluation indexes are superior to other algorithms, and the F value is up to 86%.