Cyber Threat Attack Level Detection Using Machine Learning

This study focuses on creating a machine learning-based system for identifying cyber-attacks in real time using network data, system logs, and attack history. The work presents a thorough methodology that includes data collection, preprocessing, feature selection, model training, and real-time monitoring. The system uses a variety of machine learning techniques to classify attack types, including supervised learning methods such as Random Forest, Support Vector Machines (SVM), and Neural Networks, as well as unsupervised learning methods such as clustering to detect anomalies that indicate potential threats. The model was trained and evaluated on a dataset of 477 items, reaching an accuracy of 83.33%, with 470 entries identified as carrying threats. The study effectively proved the ability to detect numerous attack types, such as SQL injection, cross-site scripting (XSS), and DDoS attacks, as well as accurately classify and prioritize threat levels. Furthermore, the system was successfully deployed in a live environment, providing real-time threat detection, automatic response capabilities, and detailed alerts to security personnel. This research resulted in the creation of an efficient and scalable cyber threat detection framework capable of both detecting existing attacks and discovering novel, previously unknown threats.