GA and SVM Algorithms for Selection of Hybrid Feature in Intrusion Detection Systems
Samira Sarvari, Zaiton Muda, Idawaty Ahmad, Mehdi Barati · International Review on Computers and Software (IRECOS) · 2015
Higher dimensionality of data that has to be analyzed for detecting attack is one of the key issues concerning intrusion detection system (IDS). This is due to the different features in the attack on network that eats up time in terms of training and prediction setups. In this study a hybrid method of Support Vector Machine (SVM) as well as Genetic Algorithm (GA) is suggested and their implementation in IDS is outlined. The suggested methods are used for reducing the number of features from 41 to 11 using KDD Cup’99 dataset. The features are classified as three priorities using GA with the most significant as the first priority and the least one as the third priority. The way in which feature distribution is done is that four features are placed in the first priority, five in the second and two in the third. The results show that the suggested hybrid algorithms, GA and SVM are able to achieve true and false positive values of 0.973 and 0.017 respectively.