Skip to Main Content
The study about static Bayesian networks with missing data has got some mature algorithms. Because of the complexity of Dynamic Bayesian Networks (DBNs), the mature methods can't be directly extended to dynamic systems. We proposed a kind of data repairing for the missing data of Discrete Dynamic Bayesian Networks (DDBNs). It incorporated network parameters, evidences and dynamic networks to predict the missing data. It's also an online repairing algorithm. Repairing algorithm was deduced in theory and verified by examples. Repaired networks make the target track and identification ability stronger. It can greatly improve the accuracy and reliability of the identification systems.