Instance Selection with Naïve Bayes to Improve DDoS Attack Classification Accuracy Using Random Forest
Aditya Putra Ramdani, Achmad Solichan, Muhammad Zainudin Al Amin, Nova Christina Sari, Basirudin Ansor, Mulil Khaira · Advances in engineering research/Advances in Engineering Research · 2024
DDoS Attack is one of the threats in a series of network systems.Attacks on a network in one unit of time can subsequently occur in a very large number of attacks.Previous research has been done to avoid DDoS attack through classification process and one of which is based on Random Forest method.The large number of attacks requires classification.In previous research, Random Forest was one way to classify DDoS attacks.The classification used is using the Random Forest algorithm.The Random Forest classification model produces an accuracy of 98.02%.This research is a preprocessing step involving Naïve Bayes instance selection which is compared with Adaboost instance selection which is expected to remove noise data due to the relatively large amount of data.With large quantities, it is hoped that this preprocessing step can get maximum results.The research also involved the Naïve Bayes and ZeroR classification methods, where the best results were using Naïve Bayes instance selection with the random forest classification method with an accuracy of 100%.