基于APDE-RBF神经网络的网络安全态势预测方法

为了提高径向基函数(radical basis function, RBF)神经网络对网络安全态势的预测精度,提出了一种基于吸引力传播(affinity propagation, AP)聚类和差分进化(differential evolution, DE)优化RBF神经网络的算法。首先,利用AP聚类算法对样本数据进行划分聚类,从而获得RBF的中心和网络的隐含层节点数;其次,利用AP聚类得出种群差异度,自适应地改变DE算法的缩放因子和交叉概率,对RBF的宽度和连接权值进行优化;同时为了避免陷入局部最优以及跳出局部极值点,对每一代种群的精英个体和种群差异度中心进行混沌搜索。通过仿真实验表明,此算法在泛化能力增强的同时,对网络安全态势也达到了较高的预测精度。

Network security situation prediction based on APDE RBF neural network

In order to improve the prediction accuracy of network security situation based on radical basis function (RBF) neural network, an optimization algorithm of RBF neural network based on affinity propagation (AP) clustering and differential evolution (DE) is proposed. Firstly, the AP clustering is used to optimize the center and the number of the hidden layer. Secondly, AP clustering is used to get the population diversity (PD), the scaling factor and the crossover probability of DE are adaptively changed with the PD for the optimized width and connection weights of RBF neural network. In order to avoid falling into the local optimum and jump out of the local extreme point, the elite individual and PD’ centers of each generation population are searched by chaotic search. The simulation results show that the APDE RBF algorithm can enhance the generalization ability, and it also has high prediction accuracy for the network security situation.