A Knowledge Graph for Network Security

In order to enable decision-making for more in-depth analysis, the network security knowledge graph can perform semantic analysis and understanding of multi-source, heterogeneous, and fragmented huge datasets. However, there are still issues with building network security knowledge graphs, such as limited datasets and ineffective entity extraction. To address this issue, this work proposes a BERT-based entity extraction model. Experiments have demonstrated that our model not only outperforms competing methods on publicly available datasets but also achieves roughly 88 percent performance on the Chinese network security dataset in terms of a variety of criteria. Finally, a network security knowledge graph is created based on knowledge extraction done on gathered Chinese network security texts.