Classification of Cognitive Patterns of Hackers Using Machine Learning

Abstract

Nowadays, computer security has become more crucial than ever. Beyond protecting digital assets, it is essential to safeguard financial institutions, companies, education, and defense sectors from increasingly sophisticated and evolving cyber threats. Addressing this challenge requires integrating various methodologies, techniques, and security tools. In this study, we utilize Honeypots, Machine Learning, and the ELK Stack, combined with an analysis of hacker psychology—understanding their motivations and behaviors—to develop more effective countermeasures. This research explores two key areas: the role of honeypots in cybersecurity and the psychological analysis of cyber attackers, examining their motivations and the tools used to measure these factors. Attack data was collected using the T-Pot Honeypot, while the Big Five Personality Traits instrument was applied to assess psychological patterns. A database was then generated, integrating this information for analysis through Machine Learning algorithms and neural networks, employing confusion matrices to compare predictions with actual data. The classification of cognitive patterns acquired through Honeypots and ML algorithms represents an emerging field that provides valuable insights into hacker behavior, enabling the development of more effective defensive strategies. Future research should focus on refining psychological assessment tools specifically designed for hackers. In our analysis, ML algorithms such as Neural Networks using a sequential model and Random Forest with 150 predictors demonstrated a strong fit for training and test datasets.