ENHANCING CYBER SECURITY THREAT DETECTION USING KNOWLEDGE GRAPHS

ABSTRACT

Cybercriminals are always changing their strategies in today's digital environment, which makes it difficult for conventional cyber-security techniques to stay up to date. This study investigates the possibilities of knowledge graph reasoning as a more flexible and advanced method to recognize and thwart network assaults in order to address this problem. This technique increases the resilience of cyber security systems by utilizing graph topologies that are infused with human-like reasoning. Three key areas are the focus of the project: knowledge graph inference methods, semantic foundations, and data preparation. The study intends to demonstrate how knowledge graph reasoning can enhance cyber-attack detection and boost the overall effectiveness of cyber security measures, including intrusion detection systems, by thoroughly examining these elements. Extensive testing has been done on the suggested strategy to confirm its efficacy in comparison to current techniques. The experiment's findings demonstrate a notable improvement over existing techniques in recognition accuracy, speed, and recall. This accomplishment makes a significant addition to the field of cyber security big data management. In the end, the research improves overall network security by laying the groundwork for the automation of network attack detection. The potential of cyber security knowledge graphs is demonstrated by their capacity to manage and rationalize information regarding cyber threats, as well as to compile and display that information. Although the majority of current research has been on the creation of a complete knowledge graph, it is still unclear how to apply the knowledge graph to address actual industry challenges in cyber attack and defense scenarios. This article provides a brief synopsis of the fundamental ideas, schema, and construction techniques of the cyber security knowledge graph. In order to facilitate future research on cyber security knowledge graphs, we also provide an overview of pertinent datasets and open-source frameworks for the information extraction and knowledge generation task. In the bulk of this research, compare and evaluate the numerous works that elaborate on the latest developments in the application scenarios of the cyber security knowledge graph. A brand-new, all-inclusive classification scheme is also created to identify the related works from nine main categories and eighteen subcategories. Finally, we have a thorough overview of numerous potential study directions based on the analyses of current research concerns.