Advancing Cyber Threat Intelligence through Machine Learning Algorithms

Advancing Cyber Threat Intelligence through Machine Learning Algorithms

연구 분야: Safety

학회: AICCONF '24: Proceedings of the Cognitive Models and Artificial Intelligence Conference

In the domain of Cyber Threat Intelligence (CTI) the enigmatic depths of the Dark Web are pivotal for the early identification of nascent cyber threats. Yet, the voluminous and sprawling data across hacker forums and illicit marketplaces pose a significant challenge, swamping analysts with a deluge of extraneous information. This paper advocates for the utilization of advanced Machine Learning (ML) algorithms as a strategic tool to distill and classify emergent threats from the Dark Web's vast data landscape. Through rigorous experiments employing a suite of ML models—Logistic Regression, Decision Tree Classification, Gradient Boosting Classifier, and Random Forest Classifier—we evaluated their performance across two meticulously curated datasets. One dataset was rich with data pertinent to cybersecurity, extracted from hacker forums and Dark Web marketplaces, while the other served as a control set. Our results illuminate the profound capability of ML algorithms to effectively navigate and filter through the data quagmire, highlighting threats with precision and thereby significantly optimizing the workflow of analysts. This automation paves the way for a more streamlined, focused approach to CTI, ensuring that cybersecurity operations remain agile and informed in the face of the dynamic and increasingly sophisticated landscape of cyber threats. Through this study, we underscore the transformative potential of ML in revolutionizing CTI methodologies, offering a beacon of efficiency and effectiveness in the ongoing battle against cybercrime.