RST-RF: A Hybrid Model based on Rough Set Theory and Random Forest for Network Intrusion Detection

Network intrusion detection system plays an important role in defending against more and more network malicious attacks. In addition, the accuracy and efficiency of the system will be significantly improved when combing different machine learning based algorithms. To this end, we build a hybrid model for network intrusion detection system which combines improved Rough Set Theory(RST) and Random Forests(RF) algorithm, named RST-RF. These two algorithms are used separately for feature selection and classification. After evaluated our RST-RF model in a public data set NSL-KDD, our method outperforms than the state-of-art researches on accuracy. As we know, our model is the first approach which combined rough set theory and random forest together for intrusion detection.