Threat assessment method based on time series forecast
Zhao Jian-ju · Journal of Zhejiang University(Engineering Science) · 2014
Aiming at the problems of transition probability getting and observational data missing in dynamic Bayesian network of threat assessment,a time series forecasting model was set up.Then the complete data set and the forward recursive algorithm were applied to parameter learning after the full data got.The threat of target was assessed based on the dynamic Bayesian network.Simulation shows that:compared to the expectation-maximization algorithm,the time series method can get higher accuracy of forecast data, have shorter learning time,increase the efficiency of the threat assessment greatly,and meet the actual operational needs.