面向感知质量保障的安全态势预测优化模型

Security situation prediction optimization model oriented to awareness quality assurance

以提高网络安全态势预测准确率为研究目标,提出一种网络安全态势预测优化模型.首先,运用层次量化分析法将多源安全态势信息进行宏观态势值量化,组成时间序列.然后,将态势值数据进行叠加归一化处理,增强数据的规律性,易于模型的建立.最后,将协方差矩阵自适应进化策略应用于支持向量机超参数的自主选择,使预测模型更加精确,并通过数据还原将预测结果加以显示.仿真实验表明该模型具有较好的预测效果.

In order to improve the prediction accuracy of network security situation,a network security situation prediction optimization model was proposed.Firstly,quantification analysis method was used to quantify multi-source security situation information for macro situation,and to construct time series.Then,the situation value data was superimposed and normalized,the regularity of data was enhanced,and the model was easy to build.Finally,the adaptive evolution strategy of covariance matrix was applied to the self-selection of support vector machines（SVM） to make the prediction model more accurate,and the prediction results were displayed by data reduction.Simulation results show good prediction effect of the proposed model.