Abstract: As cyber data attacks continue to rise, manual investigation methods are becoming increasingly inefficient, prone to errors, and time-consuming. With cyber threats evolving and attackers using similar patterns, detecting and responding to attacks in a timely manner remains a major challenge. Cyber-attacks in cyberspace aim to disrupt, disable, or take control of an organization's computing infrastructure, compromise data integrity, or steal sensitive information. The growing number of internet users and the uncertain state of cyberspace pose significant security concerns. New technological advancements and the extensive collection of big data from device sensors expose vast amounts of information, making systems more vulnerable to targeted cyber threats. Although numerous existing models and algorithms have been developed for cyber-attack prediction, there is a need for more advanced approaches that go beyond task-specific techniques.
Machine learning provides a powerful solution by framing cyber-attack prediction as a classification problem. By analyzing network datasets, supervised machine learning techniques (SMLT) can identify key patterns through variable identification, univariate and multivariate analysis, and handling missing data. A comparative analysis of various machine learning algorithms helps determine which method is most effective in predicting cyber-attacks.
Keywords: BENIGN attack, WEB attacks, SQL Injection attack, Machine learning algorithms, XSS attack, Brute Force attack, DDOS attack.
Copyright Form