Identifikasi Serangan Denial Of Service (Dos) Di Jaringan Dengan Algoritma Decision Tree C4.5

Abstract

DoS and DDoS are attacks on computer networks that flood network traffic with continuous requests. For this reason, efforts to secure computer networks and preventive measures need to be carried out by installing firewalls, IDS / IPS devices. The IDS acts as an alarm to the admin that there is abnormal activity on the network, so that the admin can take immediate preventive action. In detecting attacks, IDS uses methods or algorithms to identify anomalies that occur in the network. The algorithm is expected to be able to classify between dangerous traffic and normal traffic. Data mining is suitable to be applied in the classification of network traffic because of the large size of the data and the various types of attacks. The C4.5 decision tree algorithm is expected to be able to be used in the traffic classification process with the aim of identifying DoS attacks. The results of the trial with dataset testing, C.45 yielded an accuracy of 90,68% in classifying traffic for the identification of DoS attacks, and yielded an accuracy of 74,99% in classifying all types of traffic. The Naïve Bayes algorithm is used as a comparison, the accuracy is 86,56% in classifying DoS attack identification traffic, and produces an accuracy of 69,50% in classifying all types of traffic. The C4.5 algorithm is superior in terms of accuracy but takes longer to build the model than the Naïve Bayes algorithm.