Improved Machine Learning-Based Model for Network Traffic Anomaly Detection

This study presents an improved technique for developing an effective anomaly detection for network traffic using the Random Forest classifier, optimized with hyperparameter tuning. The classifier was trained and evaluated using 5-fold cross-validation, resulting in optimal hyperparameters and achieving an impressive accuracy of 97.8%, and 98% for precision, recall, and F1-score, underscoring the model's balanced and reliable classification of normal and attack traffic. This study highlights the significant potential of Random Forest classifiers in intrusion detection and sets the stage for future research, which could involve integrating additional data sources, including deep learning techniques, and enhancing real-time detection capabilities to adapt to the ever-changing attributes of cyber threats.