CYBER-ATTACK PREDICTION USING MACHINE LEARNING ALGORITHMS

Abstract

Cybersecurity is increasingly vital as cyberattacks grow in both frequency and complexity. This study investigates the use of machine learning models—XGBoost, Decision Trees, and Support Vector Machines (SVM)—for predicting cyber threats based on historical data and network behavior. By analyzing traffic patterns, system logs, and user activities, these models detect potential threats with high accuracy. XGBoost enhances performance through boosting, Decision Trees provide clear threat classification, and SVM excels in handling complex attack patterns. Together, these techniques enable more accurate and proactive threat detection, improving system resilience and reducing false positives.