Evaluation on Prediction Algorithms based on Computer Virus Trend
Jie Ma · Information Security and Communications Privacy · 2011
Forecast on the development trend of computer virus can help the judgment of network security situation for proper precautionary preparations. Currently forecast on the development trend of computer virus is still in exploratory stage,the forecast results of traditional method in this regard are still not clear. Various prediction algorithms based on actual computer virus are compared,including moving average,trend forecasting,seasonal changes forecasting,and linear regression and so on. The analysis on different time spans,different types of malicious code data indicates that the linear regression and seasonal change forecasting algorithm have better performance in the periodic computer virus log data.