Abstract: In the face of rapidly advancing cybersecurity threats, Intrusion Detection and Prevention System (IDPS) have established themselves as critical tools for warding off harmful activities against a network. Based on this consideration, this review tracks the development and impact of Machine Learning and Deep Learning strategies as associated with IDPS, focusing particularly on their ability to enhance detection performance. We have Surveyed various Intrusion Detection and Prevention System Datasets for assessing their effectiveness in detecting network intrusions. More importantly, it focuses on critical datasets and talks about the pros associated with them, such as better detection capability and their flexibility toward ever-evolving threats, but failed to fight some limitations like increased computational complexity and complex real-time traffic management. This survey gives an overview of the evolution and effectiveness of "Machine Learning and Deep Learning" techniques in advancing IDPS, addressing major concerns over issues of scalability, false positive rates, accuracy, Recall, Precision, F1 Score and overall system efficiency, with an aim to improve the fairness and reliability of intrusion detection and prevention system mechanisms.
Keywords: Intrusion Detection and Prevention System, Machine Learning, Deep Learning, Network Security, Random Forest, Support Vector Machine, Convolutional Neural Networks, Cybersecurity, Anomaly Detection, False Positives, Real-time Traffic, Scalability, Detection Accuracy.
|
DOI:
10.17148/IJARCCE.2024.131019
