Machine Learning Approaches for DDoS Attack Detection: Naive Bayes vs Logistic Regression

DDoS assaults are one of the biggest cyber threats. These attacks flood servers with malicious traffic to prevent them from serving legitimate users. In this case, the server's bandwidth and buffer size are severely slowed or depleted, disrupting service. Researchers have devised a mathematical approach to analyze distributed denial-of-service assaults. This model helps understand DDoS attack patterns. Mathematical representations help researchers understand assault processes and create effective defenses. Machine learning techniques discern DDoS assaults from network traffic. The work uses Logistic Regression and Naive Bayes to detect attacks and characterize typical conditions. KDDcup, a dataset commonly used for network security experiments, trains these algorithms. The researchers used Weka, a data mining platform with many data analysis and machine learning tools. Using the KDDcup dataset, they introduced and tested machine learning algorithms to reliably detect DDoS assaults and distinguish them from regular network traffic. The taught algorithms were then tested in real-world circumstances. The study's findings were compared to machine learning-based denial-of-service attack detection studies. The researchers compared algorithm performance to assess their proposed approach and identified areas for improvement.