A model for multi-class intrusion detection using the dragonfly feature selection and Random forest Algorithm on the CICIDS-2017 dataset

With the growth of information technology knowledge and the expansion of its applications, the development of new security models and the analysis and design of appropriate methods for detecting intrusion into networks and systems has become particularly important. In this research, a model for intrusion detection called ID2F based on feature selection using dragonfly algorithm and random forest classification has been proposed and proposed. The proposed method is a multi-class method, in other words, in addition to detecting intrusion, it also determines the type of attack. In this study, two completely different datasets, CICIDS-2017 and KDD-CUP99, were used for analysis to evaluate the performance of the method with a separate dataset. The problem is implemented with different algorithms and the best algorithm is selected as the proposed method. The accuracy value in the proposed method in the CICIDS2017 dataset is 99.83 and for the KDD-CUP99 dataset is 99.85. In addition, the results of the research have been compared with several other methods proposed by previous researchers, and this comparison shows that the proposed method is more accurate than most machine learning methods and its implementation time is better.