Intrusion Detection System using Genetic Algorithm and K-NN Algorithm on Dos Attack

Abstract

Intrusion Detection is the process of monitoring and identifying activity on a host or network to prove whether the host or network has been successfully attacked or is still an attempt at aggression. Intrusion Detection System (IDS) helps monitor the network based on various anomalies (unusual events) that can indicate threats of hacker aggression, malware, or vulnerabilities in a system. IDS will monitor and provide a warning of whether an activity is classified as malicious or not. Furthermore, IDS will organize it into several strata levels of risk. This is very helpful for prioritizing any activity anomalies that require more attention and handling. This study analyzed the IDS process with a selection feature using genetic algorithms and classification using the KNN algorithm and KDD99 as the dataset. By selecting the best features from 41 to 18, the scenario in this study gets an average training data accuracy of 99.98% and testing data of 97.52% in the parameters K = 5 and K = 7.