一种基于似然BP的网络安全态势预测方法

Method of Network Security Situation Prediction Based on Likelihood BP

态势预测是网络安全态势感知的高级阶段。为了解决依赖于专家赋予权值、缺乏自学习的态势数据处理方法在复杂网络系统中的局限,提出了一种基于似然BP的网络安全态势预测方法,将BP神经网络引入态势预测领域,并用极大似然误差函数代替传统的误差函数,通过态势评估模型建立的态势序列作为训练输入序列,在反向传播过程中实现对指定参数权值的自学习调整,该方法能充分利用网络越复杂、粒度越细、效率就越高的特点,实验表明了该方法具有较好的态势预测效能,为网络安全态势预测提供了一种新的解决途径。

Situation prediction is the advanced stage of network security situation awareness. For purpose of resolving the limitations of depending on experts giving weight, lacking of self-learning on data processing in complex network system, a method of network security situation prediction based on likelihood BP was proposed. The BP neural network was introduced to the situation prediction area, and the traditional error function was replaced by the maximum likelihood error function. The situation sequences established through the situation assessment model were used as the training input sequences, and the self-learning adjustment of the appointed parameters＇ values was implemented in the process of back propagation training. The new method can make full use of the characteristics of the network more complex, finer grain size, the higher the efficiency. Experimental results show that the method has good performance of situation prediction,and provides a new solution for network security situation prediction.