With the growth of the internet, network attacks have increased severely in a substantial number in the last few years. Therefore, Intrusion Detection Systems (IDSs) have become a necessary addition to the information security of most organizations. An IDS monitors a network or a single host looking for suspicious activity and reports them. Many intrusion detection types of research have focused on the feature selection because some characteristics are irrelevant or redundant which result in a lengthy detection process and degrades the performance of IDS. For this purpose, we have used in this work an algorithm based on Information Gain technique. This algorithm selects an optimal number of features from NSL-KDD Dataset. In addition, we have combined the feature selection with a machine learning technique named Support Vector Machine (SVM) using Radial-basis kernel function (RBF) and a Particle Swarm Optimization algorithm to optimize the parameters of SVM for effective classification of the dataset. We have also compared the proposed method and other methods. Tests on the NSL-KDD Dataset have proved that our proposed method can reduce the number of features and obtain good results in terms of accuracy, attack detection rate and false positives rate, even for unknown attacks.
展开▼