A smart network intrusion detection system based on network data analyzer and support vector machine

• Main Authors: Babatunde O.S., Ahmad A.R., Mostafa S.A., Foozy C.F.M., Khalaf B.A., Fadel A.H., Shamala P.
• Other Authors: 57219411278
• Format: Article
• Published: World Academy of Research in Science and Engineering 2023

Because of the critical interest for viable IDS in networks security, the researchers are trying to recognize enhanced methods. This work shows how the KDD dataset is exceptionally helpful for testing distinctive DDoS classifiers. Conclusively, there are two principal ways to reduce the classification complexity and improve the DDoS attack detection accuracy by using nonlinear Support Vector Machine (SVM)s: (1) reducing the number of support vectors; (2) simplifying the classification process for special kernels. This paper proposes a Smart Intrusion Detection System (SIDS) that integrates a Network Data Analyzer (NDA) and SVM to reduce the computation iterations needed by the SVM by eliminating the presumed attack types before performing the classification process. Reduction in data can also serve as a way to increase speed and reduce time in computations. Also, it enhances performance evaluation as 3 types of attack are easier to evaluate than 4 types especially where the 4th type is dominant in the analyzed datasets (the case of DDoS attack being about 79% of the total dataset). As experimented, the proposed Smart Intrusion Detection System method has shown a way in dataset reduction by simply eliminating the DDOS attack types with the same amount of data as compared to Batch 2. Batch 1 serves as a control experiment as indicated by its good performance evaluation measurements. � 2020, World Academy of Research in Science and Engineering. All rights reserved.