Intelligent SSH Attack Detection Model Using K-Clique Clustering and Reinforcement Learning

The research aims at developing a system that applies graph-based clustering of log data-more specifically, k-clique percolation to detect and analyze SSH brute-force attacks. The process involves preprocessing SSH data, constructing an IP interaction network graph, and then applying clustering algorithms to identify patterns that are indicative of malicious behavior. Furthermore, integration of a reinforcement learning model, specifically the q-learning algorithm, with cyber security will strengthen the system's predictive abilities and enable it to adapt continually to new attack vectors. The aim is to develop a powerful tool for tracking, evaluating, and preventing SSH brute-force attacks.