Advanced Threat Recognition: Harnessing Machine Learning for Intrusion Detection
Ambalika Dey · International Journal for Research in Applied Science and Engineering Technology · 2024
Conventional intrusion detection systems (IDS) are frequently inadequate to combat complex cyber threats in the changing cybersecurity landscape. To create a reliable intrusion detection system, this work investigates the use of machine learning, more especially the RandomForestClassifier method. The study highlights that in order to maximize model training, careful data preprocessing—including feature selection, label encoding, and scaling—is essential. It also uses Optuna for hyperparameter adjustment to improve the classifier's performance. As a scalable and dependable response to contemporary cybersecurity issues, the findings show how effective machine learning is in identifying sophisticated attacks. The outcomes show a high detection accuracy and resilience to complex intrusions, confirming machine learning's potential as a scalable and reliable answer to today's cybersecurity problems