Network Anomaly Detection Utilizing Machine Learning Methods

The rapid expansion of technology and the growing dependence on networked systems have elevated network security to a paramount concern for both individuals and organizations. In the face of ever-evolving and increasingly sophisticated cyber threats, conventional rule-based intrusion detection systems often struggle to maintain pace. Machine learning can offer potent solutions for identifying anomalies in network traffic and potential security breaches. While signature-based methods are commonly employed for attack detection, they are ineffective at countering zero-day attacks. This article discusses an alternative approach, the anomaly-based method, which is adept at identifying network attacks, including zero-day attacks. The primary objective of the research is to employ machine learning algorithms for the detection of anomalies within computer networks. To achieve this goal, the versatile CICIDS2019 database is used. The criteria are chosen from the test dataset using the random forest regression algorithm. Seven distinct machine learning algorithms are employed, and the outcomes are assessed using performance metrics including precision, recall, and F-measure. The algorithms exhibited a strong performance in line with prior studies, demonstrating their effectiveness.