Intrusion Detection Methodologies Based on Machine Learning:Feature Selection, Datasets, Performance Measures and Results

Due to the increased need for Internet access in variousindustries and the substantial applications relating to computersthat have been developed recently, cyber-attacks on the Internethave increased, as has the challenge of cyber-security. Thisnecessitated the adoption of an intrusion detection system. Ithandles traffic data in order to keep track of the network and thedevices linked to it in order to spot any malicious activity orattacks on websites or online applications. It has becomeimportant to use an intrusion detection system to inspect datatraffic within the network to assure its confidentiality, integrity,and availability. Despite the researchers' best efforts, IDS stillhas difficulties boosting detection accuracy while lowering falsealarm rates. Therefore, many machine learning techniques havebeen introduced to intrusion detection systems. This paperpresents an overview of the recent literature related to techniquesfor detecting intrusion and cyber-attacks using machine learningalgorithms. And the challenges we face in detecting intrusion inthe network and applying machine learning also shed light on thedata sets used in training the proposed models.