A Review on Intrusion Detection Using Machine Learning Techniques

Abstract

An essential tool for monitoring and identifying intrusion threats is the intrusion detection system (IDS). As a result, intrusion detection systems monitor network traffic heading through computer systems to detect for malicious activity and recognized dangers, and send alerts. With a focus on datasets, ML methods, and metrics, this study tries to analyse recent IDS research using a Machine Learning (ML) approach. To make sure the model is suitable for IDS application, dataset selection is crucial. The efficiency of the ML method can also be impacted by the dataset structure. As a result, the choice of ML algorithm depends on the dataset's structure. Metric will then offer a quantitative assessment of ML algorithms for a given dataset. In addition True Positive Rate (TPR), False Positive Rate (FPR) and accuracy, are the three metrics for IDS performance evaluation that are most frequently utilized. This is understandable given that these metrics offer crucial cues that are crucial to IDS performance. A clear path and direction for future study has been provided by the discussion and comparison of the results from various works.