基于PSO_SVR的网络安全态势预测方法

NETWORK SECURITY SITUATION PREDICTION METHOD BASED ON PSO_SVR

针对网络安全态势感知中的态势预测问题,提出一种基于PSO_SVR的网络安全态势预测方法。该方法将支持向量回归机(SVR)嵌入到粒子群优化算法(PSO)的适应度计算过程中,利用PSO算法的全局搜索能力来优化选取SVR的参数,在一定程度上提升了SVR的学习能力和泛化能力。仿真实验表明,通过与已有的其他预测方法作对比,该方法具有更好的预测效果。

To address the situation prediction issue in network security situation awareness,we present a PSO-SVR-based network security situation prediction method. This method embeds support vector machine for regression（SVR） into the fitness calculation process of particle swarm optimisation（PSO） algorithm,and makes use of global searching capability of PSO to optimise the selection of SVR parameters. To some extent,this enhances the learning ability and generalisation ability of SVR. Simulation experiments show that this method has better prediction effect in comparison with other existing prediction methods.