基于传统象群优化-粒子群优化-支持向量机的网络入侵检测

Research on network intrusion detection based on particle swarm optimizationelephant herding optimization-support vector machines

为提高网络入侵检测的准确率,提出一种改进象群优化算法的支持向量机(support vector machines, SVM)入侵检测模型。针对传统象群优化算法(elephant herding optimization, EHO)在全局搜索上的劣势,提出采用粒子群优化算法(particle swarm optimization, PSO)对EHO进行优化的方法,以提高全局搜索能力;基于以上改进,对SVM模型参数进行寻优,并以寻优模型为基础,构建网络入侵检测模型;以KDD CUP99数据集为基础,对比改进的PSO-EHO-SVM模型与EHO-SVM,PSO-SVM和飞蛾扑火优化(moth-flame optimization, MFO)-SVM在最优适应度计算、入侵检测分类准确率和算法稳定性方面的优劣。仿真结果表明:PSO-EHO-SVM迭代次数最少,分类准确率更高,且稳定性更强,证明改进后的SVM模型更具有优势。

To improve the accuracy of network intrusion detection, an intrusion detection model by using the support vector machines(SVM) based on improved image swarm optimization algorithm is proposed. Aiming at the disadvantage of traditional image swarm optimization algorithm in the global search, the particle swarm optimization(PSO) is proposed to optimize the elephant herding optimization(EHO)to improve the global fast search ability;Based on the above improvements, the parameters of SVM model are optimized, and the network intrusion detection model is constructed based on the optimization model. Using the KDD CUP99 data set, the advantages and disadvantages of the improved PSO-EHO-SVM model and EHO-SVM, PSO-SVM and MFO-SVM in optimal fitness calculation, intrusion detection classification accuracy and algorithm stability are compared. The results show that the PSO-EHO-SVM method has the least iteration times, higher classification accuracy and stronger stability, and is proved that the improved method is feasible and has more advantages.