Assessing Intrusion Detection Process Using ML Techniques: Issues, Options, and Potential Future Directions

In today's society, the world of computer networks is rapidly expanding and has emerged as an essential means of information sharing. However, given the widespread use of networks for information sharing, data security has become even more critical. In spite of the fact that a variety of methods have been developed to safeguard data, intruders continue to develop novel strategies for breaching network security as new technologies emerge. R2L and U2R are two types of attacks that are considered to be uncommon due to their rarity in networks. However, because they allow intruders to gain access, these attacks can be very dangerous for the hosts they target. The current security system is limited in its ability to accurately, precisely, and without false alarms detect and prevent such attacks. A ton of examination is being finished to upgrade network security. This article discusses various measurable outcomes and itemized correlation with various strategies. Furthermore, with and without the use of feature selection techniques, we compare and contrast the attack detection capabilities of a number of existing approaches in this paper. New researchers may take benefit from these evaluative efforts to determine the most effective strategy for detecting rare class attacks and the significance of incorporating feature selection schemes. NSL-KDD datasets are used to test and investigate an assessment.