INTRUSION DETECTION OF IMBALANCED NETWORK TRAFFIC BASED ON MACHINE

Abstract

The security of today's wireless networks is under constant danger from a growing 
number of faults, vulnerabilities, and assaults due to the fast increase of these types of 
problems. Network security is becoming an increasingly essential topic as a direct result of the 
fact that computer networks and applications are constructed on different platforms. Operating 
systems that are both complicated and costly are more likely to have flaws in their security. 
efforts to breach security, completeness, or availability are referred to as "intrusions," and the 
word "intrusion" describes such efforts. An intrusion detection system (IDS) is useful for 
detecting flaws in network security as well as unusual activity. Despite the fact that it is 
sometimes seen as premature and not as an ultimately complete technique of combating 
intrusions, the development of technology for detecting intrusions has been a sector that has 
had significant growth in recent years. The completion of this assignment has been elevated to 
the status of a top priority by both security professionals and network administrators. 
This indicates that even more secure systems are unable to totally take its place. IDS is able to 
anticipate future intrusions based on intrusions that have already been identified since it uses 
data mining to identify intrusions. In this study, an exhaustive literature evaluation on the use 
of data mining techniques for IDS is offered. In the first step of this process, we will examine 
data mining techniques for the purpose of identifying intrusions using real-time and benchmark 
information. This article provides a comparison of several approaches of detecting intrusions 
in a network, along with an analysis of the benefits and drawbacks of each strategy. Within the 
scope of this research, we suggest many methods that might enhance network intrusion 
detection. 

,