Machine Learning Algorithms for Monitoring & Detecting Cyber Attacks

In our highly connected world, protecting digital equipment from cyberattacks has become more important than ever. This study explores how machine learning techniques can be implemented to identify cyberattacks by analyzing the traffic patterns. The dataset utilized in this study includes essential network packet attributes like packet size, source IP and destination IP, protocol information, and TCP/UDP port values. Four distinct classifiers were evaluated for binary classification to distinguish cyberattack activity from normal network behavior. Kneighboursclassifier, Logistic regression, AdaBoost, and Random Forest were chosen because of their effectiveness for this type of task and their frequent application in the cybersecurity field. The performance was evaluated using several metrics, including F1 Score, recall, accuracy, precision, and confusion matrix. The result of the study highlights the advantages and the drawbacks of various machine learning algorithms in identifying intrusions, contributing to the development of more robust network security solutions.