Constructing a Cyber Security Knowledge Graph: a Comprehensive Approach to Navigate Digital Threats

Abstract

In this thesis we present the undertaken approach of building a single, easily queryable cyber security knowledge graph by interlinking heterogeneous publicly available information on attack patterns, weaknesses, vulnerabilities, numbering authorities and related exposed platforms that otherwise would be sparsed in multiple isolated scopes, ultimately resulting in an increased effort - defined in terms of time and overall searchability - for the end-users. The resulting monolithic dataset allows, thanks to its magnitude and capillarity, the execution of sophisticated queries that can potentially provide new insights and valuable perspectives that weren't possible before. Alongside with the standard technology-related query interfaces generally reserved to developers, which in this instance refers to Neo4j Browser and Bloom, a Single Page Application (SPA) was also created to provide a deployable online general-user endpoint for interacting with the newly constructed knowledge graph.