基于混合核函数的LSSVM网络入侵检测方法

摘要

针对常规网络入侵检测算法检测率低、误报率高以及检测效率低下等问题,在此使用基于混合核函数的最小二乘支持向量机作为网络入侵检测模型的核心算法,使用粒子群优化算法对最小二乘支持向量机的各个参数进行优化.使用著名的KDD CUP99数据库中的部分数据样本对网络入侵检测模型进行训练和测试,以验证所提出网络入侵检测方法的性能.测试实验结果表明,提出的基于混合核函数的PSO-LSSVM算法具有更好的检测性能,提高了检测系统的检测率.%Since the conventional detection algorithm of network intrusion has low detection rate,high false positive rate and low detection efficiency,the least squares support vector machine (LSSVM) algorithm based on hybrid kernel function is taken as the core algorithm of the network intrusion detection model,and each parameter of the LSSVM is optimized by using particle swarm optimization(PSO)algorithm. The network intrusion detection model was trained and tested by partial data sam-ples in famous KDD CUP99 database to verify the performance of the proposed network intrusion detection method. The test re-sults show that the PSO-LSSVM algorithm based on hybrid kernel function has better detection performance,and can improve the detection rate of the detection system.