Application of Improved Fireworks Algorithm Optimized SVM in Intrusion Detection
He Li, Xiang Ji, Jingmei Li · Proceedings of the 2018 International Conference on Network, Communication, Computer Engineering (NCCE 2018) · 2018
In order to further improve the effectiveness and accuracy of intrusion detection and avoid the security risks brought by network attacks, this paper proposes an intrusion detection strategy based on improved fireworks algorithms-TFWA-SVM.This algorithm first optimizes the fireworks algorithm, constrains the initial fireworks position in the fireworks algorithm, avoids the waste of computing resources caused by the initial concentration of the fireworks.At the same time, it uses the fitness function stretching technology to make the algorithm have more superior global exploration capabilities.The improved fireworks algorithm is applied in the selection of SVM penalty factor and kernel function parameters, and then the powerful classification ability of SVM is used to classify the data packets in the network.Using the BPNN, SVM, FWA-SVM and TFWA-SVM to simulate the KDD99 data set.The experimental results show that TFWA-SVM has obvious advantages in convergence speed and classification accuracy.It can improve the quality of intrusion detection to a certain extent and has extensive research value.