Intrusion Classification Detection Model for SDN based on Optimized TSO and DT
Zhengbin Qin, Hui Zhi Xu, Lang Huang, Jin Yang · 2023
For the attack and anomaly problems in Software-Defined Networking (SDN), an SDN network intrusion classification detection model based on optimized Tuna Swarm Optimization (TSO) algorithm and Decision Tree (DT) is proposed. First, feature selection based on multi-strategy Improved Tuna Swarm Optimization (ITSO) algorithm is adopted to optimize the dataset and reduce feature redundancy and dimensionality. Second, multi- classification algorithm of DT is optimized to improve its performance and accuracy. Finally, a model based on optimized TSO and DT is constructed for intrusion classification detection in SDN networks. The classification accuracy rate is 0.9582 on the InSDN dataset with 10% data volume, and the score on three datasets with 5,000 data points is above 0.95. Experimental results show that the proposed SDN intrusion classification detection model has good detection performance.