Information security risk estimation based on auto regressive integrated moving average and relevance vector machine
Wei Jie-mi · Jiguang zazhi · 2015
The single estimation models of information security risk have low accuracy n problems,comprehensive utilization of auto regressive integrated moving average and relevance vector machine advantage,this paper puts forward a information security risk estimation model based on ARIMA-RVM.Firstly,the historical time series data are collected for the information system security status,and t local mean decomposition algorithm is used to decompose time series data and generate some component,secondly,the low-frequency components are modeled by auto regressive integrated moving average while frequency change value is large,volatile component are modeled by relevance vector machine,finally,the estimate results of components are added to get final results of information security risk.Experimental results show that the proposed model can accurately describe the change trend of information system,improve the information security risk estimation accuracy and has higher practical application value.