A Deep Learning-Based Knowledge Discovery Framework for Evolving Cyber Threat Landscape

As cyberattacks grow more frequent and complex, organisations need faster ways to understand emerging digital threats. Much of today’s early cyber threat information appears in unstructured online sources such as social media, blogs, and forums, which are difficult to analyse at scale. This thesis develops new text mining methods for cyber threat knowledge discovery, enabling the automatic identification, organisation, and analysis of large volumes of online security information. By transforming raw online discussions into structured and actionable knowledge, this research helps cybersecurity professionals detect emerging risks earlier and respond more effectively in a rapidly evolving digital landscape.