Intrusion detection in computer networks through a hybrid approach of data mining and decision trees
Tayebeh Rouhani Nejad, Mohammad Ebrahim Shiri Ahmad Abadi · 2014
Increasing dev elopment of information and a trend towards using digital resourc es has posed a challenge to information sec urity. The most important issue is to protec t c omputer systems against infiltrators and misusers. To protec t computer sy stems and networks against infiltrators, sev eral approaches have been designed called intrusion detection approac hes. The purpose of intrusion detection approac h is to detec t any unauthorized ac tivities, misu ses, and damage to computer systems and networks by internal users or external attackers. Intrusion detection sy stems are one of the major fac tors of sec urity substruc tures for several organizations. These sy stems are models, and hardware and software patterns that automatize proc esses. They notify user as an alarm. Sometimes these alarms are correct and sometimes inc orrect. To avoid these alarms, data mining-based intrusion detection sy stems are used. In this paper, we suggest a hybrid approac h of data mining with feature reduc tion techniques c ontaining 41 features and decision tree algorithms to improve performance (97.19%).