Enhancing Cybersecurity: A Pipeline Approach for Efficient Machine Learning Intrusion Detection

The escalating commonness of severe cyberattacks on a global scale has advanced the critical need for integrating machine learning into cybersecurity practices. Seven models, trained on a reduced feature inSDN dataset, aim to detect malicious SDN traffic. Testing involves diverse traffic types generated with Pyshark and CICFlowMeter. The study achieves its goals by creating effective intrusion detection models, developing new test datasets, implementing the pipeline, and validating it through testing in the real-time practical network in the cybersecurity lab. The use of double-feature selection is pivotal. Overall, this research accelerates the deployment of supervised machine learning in intrusion detection, strengthening cybersecurity against evolving threats.