李方伟, 罗嘉, 朱江, 张海波. 一种基于混合核函数PSO_SVR的网络安全态势预测方法[J]. 微电子学与计算机, 2015, 32(12): 110-115.
引用本文: 李方伟, 罗嘉, 朱江, 张海波. 一种基于混合核函数PSO_SVR的网络安全态势预测方法[J]. 微电子学与计算机, 2015, 32(12): 110-115.
LI Fang-wei, LUO Jia, ZHU Jiang, ZHANG Hai-bo. A Method of Network Security Situation Prediction Based on Hybrid Kernels PSO-SVR[J]. Microelectronics & Computer, 2015, 32(12): 110-115.
Citation: LI Fang-wei, LUO Jia, ZHU Jiang, ZHANG Hai-bo. A Method of Network Security Situation Prediction Based on Hybrid Kernels PSO-SVR[J]. Microelectronics & Computer, 2015, 32(12): 110-115.

一种基于混合核函数PSO_SVR的网络安全态势预测方法

A Method of Network Security Situation Prediction Based on Hybrid Kernels PSO-SVR

  • 摘要: 为了对错综复杂的网络安全形势做出可靠的预测,提出了一种基于混合核函数PSO_SVR的网络安全态势预测模型.本模型针对基于传统支持向量机(SVR)的网络安全态势预测模型精度不够高,其核函数的选择及参数的设定没有统一标准的情况,构造了一种兼顾插值能力和外推性能的混合核函数.并引入粒子群算法(PSO)对基于混合核函数的SVR进行参数寻优,有效地提高了SVR预测能力.通过仿真实验表明,该模型相比与传统的网络安全态势预测方法,预测精度上更有保障.

     

    Abstract: In order to predict the complicated network security situation reliably, a hybrid network security situation predictive model based on kernel function PSO_SVR is proposed. To solve the problem that the accuracy of prediction is low and there is no uniform to set the parameters, we construct a hybrid kernel function whose interpolation and extrapolation performance is good. The particle warm optimization (PSO) based on hybrid kernel function is introduced to search the optimization parameters. The simulation results show that the model compared with the traditional network security situation prediction method is more secure than in the prediction accuracy. The predictions are more scientific and reliable.

     

/

返回文章
返回