萤火虫群算法优化高斯过程的网络安全态势预测

发布时间：2018-06-28 04:30

本文选题：态势预测 + 高斯过程　；参考：《系统工程与电子技术》2015年08期

【摘要】：针对共轭梯度法获取高斯过程超参数存在迭代次数难以确定及预测不精准等问题,提出一种萤火虫群算法优化高斯过程的预测方法,并将其应用于网络安全态势预测研究。采用萤火虫群优化算法对高斯过程超参数进行智能寻优,建立基于高斯过程回归的网络安全态势预测模型。实验结果表明新方法的平均相对预测误差较共轭梯度法、粒子群优化算法和人工蜂群优化算法分别降低了近29.46%、10.37%和4.22%,且新方法收敛较快。另外,分析对比了3种单一类型和2种复合类型的协方差函数对高斯过程预测的影响,实验结果表明采用神经网络与有理二次的复合协方差函数(neural network and rational quadratic composite covariance function,NN-RQ)的平均相对预测误差较其他4类协方差函数降低了1.65%~7.51%。
[Abstract]:In order to solve the problems such as the number of iterations and the inaccuracy of prediction in obtaining superparameters of Gao Si process by conjugate gradient method, this paper presents a prediction method for optimizing Gao Si process by firefly swarm algorithm, and applies it to the research of network security situation prediction. The Gao Si process superparameters are intelligently optimized by firefly swarm optimization algorithm, and the network security situation prediction model based on Gao Si process regression is established. The experimental results show that the average relative prediction error of the new method is 29.46% and 4.22% lower than that of the conjugate gradient method, particle swarm optimization algorithm and artificial bee colony optimization algorithm, respectively, and the convergence of the new method is faster. In addition, the effects of covariance functions of three single types and two complex types on the prediction of Gao Si processes are analyzed and compared. The experimental results show that the average relative prediction error of the composite covariance function (neural network and rational quadratic composite covariance functionNN-RQ using neural network and rational quadratic is 1.65 / 7.51 lower than that of the other four covariance functions.
【作者单位】：空军工程大学信息与导航学院;空军工程大学空管领航学院;
【基金】：国家自然科学基金(61201209)资助课题
【分类号】：TP393.08

【参考文献】