基于GMM改进的信息系统安全态势实时预测研究

RESEARCH ON REAL-TIME FORECAST OF SECURITY POSTURE OF INFORMATION SYSTEM BASED ON IMPROVED GREY-MARKOV CHAIN

现有安全态势预测方法由于消耗较大和耗时较长而造成预测效果不佳,考虑到信息系统的结构复杂、信息交互频繁等特点,依据典型灰色模型GM(1,1)的消耗低、样本小、适用性强、短期预测效果好等特点对信息系统进行实时的安全态势预测。同时针对GM(1,1)模型的随机波动性小的问题,结合马尔可夫(Markov)链适用于随机波动较大的特点,提出一种以灰色GM(1,1)为预测原型,用马尔可夫链对GM(1,1)预测模型进行误差修正的实时信息系统安全态势预测模型。实验结果表明,在信息系统安全态势预测方面,该模型能够较准确地预测安全态势的总体趋势,且预测精度高于原灰色-马尔可夫模型的精度。

Aiming at the problem of bad prediction result because of large consumption and time consuming in the existing security posture prediction methods,the information system is real-time forecasted in security posture according to the characteristics of low consumption,small sample,stronger applicability and excellent performance of short-time forecasting of the typical gray GM（1,1） model,considering the problems of large-scale,complex structure and frequent information exchange of information system. Meanwhile,aiming at the problem of low stochastic volatility of GM（1,1）model,an improved GM（1,1） model is built,which combines with the characteristic of larger random fluctuation of Markov chain applies. Thus,using gray GM（1,1） as a prototype of forecast,a real-time forecast of information system security posture is proposed by utilizing the Markov chain to modify the improved GM（1,1） model. The experimental results show that the model is able to predict the overall trend of security posture accurately and is better than the previous Grey-Markov model in the forecast of information systems security posture.