Network Security Threat Detection System Based on Knowledge Graph

Due to the growing sophistication and heterogeneity of information technology, cybersecurity has become a critical issue in the contemporary era. Existing cybersecurity frameworks and databases provide structured methodologies for threat intelligence and vulnerability assessment. However, they severely lack multi-source heterogeneous data integration, dynamic threat environment adaptation, and a natural language interface for intuitive access. This work presents a knowledge graph-based cybersecurity threat detection system that addresses these challenges by (a) enriching data from heterogeneous sources into a semantic framework, (b) building a dynamic vulnerability scoring model that evaluates threats using contextual and attack-chain information, and (c) providing an intuitive natural language querying interface. As an example of the system’s practical utility, we demonstrate dynamic scoring with a real-world threat scoring case study, highlighting its ability to analyze and rank vulnerabilities and risks accordingly. This research bridges the gap between theory and practice by integrating knowledge graphs, dynamic scoring models, and advanced querying techniques to overcome cybersecurity threat detection and vulnerability assessment challenges.