查看文章

In this paper, five primary classification methods with three feature selection strategies have been implemented to classify the network attacks using NSL-KDD dataset. These methods are (J48 decision tree, Support Vector Machine (SVM), Decision Table (DT), Bayesian Network and Back Propagation Neural Network). The feature selection strategies are (Correlation base feature selection (CFS), Information Gain (IG) and Decision Table). Several experiments have been implemented to obtain good results using the training and testing NSL-KDD within general attack (Normal and Anomaly). These were carried out using four attack types: Denial of Service attack (DOS), User to Root attack (U2R), Remote to Local attack (R2L) and Probing attack. J48 classification method with information gain feature selection gives the best results (80.3%) using testing dataset and (93.9%) as an accuracy training dataset.