<![CDATA[ABSTRACT Intrusion Detection System is the ability possessed by hardware or software that serves to detect suspicious activity on the network and analyze and search in general. The purpose of this study is to classify attack detection on the Intrusion Detection System using the C.45, Naïve Bayes and K-NN algorithms to see how big the attack is. The benefits gained in this study are as a test and learning material in analyzing, classifying attacks so that they can prevent and minimize attacks to users. To overcome this problem, this study uses the C.45 algorithm, Naïve Bayes, K-NN, K-NN algorithm produces an accuracy rate of 82.58%, Recall 81.73% and Precision 84.11% while the Naïve Bayes accuracy 96.91%, Recall 97,45% and Percision 96.18% and the algorithm produces an optimal value of C.45 accuracy 97.80% Recall 98.18% and Precision 97.60%. On the attribute (attack) which has the number of classes or normal labels, dos, probes, r21. The results of the lowest K-NN algorithm are caused or normal to be considered yes(an attack) which should be No(no attack)and the C.45 algorithm attribute(attack) normal, dos, probe and r21, normal(no attack), yes(the presence of an attack) is optimal in the classification of attack detection data on Intrusion Detection System(IDS). Keywords: Data Mining, C.45, Naïve Bayes and K-NN, Intrusion Detection System(IDS)]]>
Copyrights © 2021
