Dynamic Bayesian network threat assessment for warship formation: a data analysis method

Abstract

In the target threat assessment of maritime formation air defence, the observation data are easy to be missing, and existing data analysis methods are difficult to carry out dynamic assessment in time series. In order to solve these problems, a data analysis method about threat assessment is proposed, which is based on discrete dynamic Bayesian networks (DDBN) and the utility theory. Firstly, the data characteristics of the target threat assessment are analysed, and a two-stage dynamic Bayesian network structure evaluation system is constructed. Secondly, the continuous variable in the network structure is transformed into a discrete variable, which can avoid the repeated calculation caused by the continuous change of the node threat attribute value in a small range. Then, the prior probability of the credibility of the uncertainty node to make the Bayesian network parameters more realistic, and the utility theory is introduced to carry out the threat ranking. Finally, the simulation results show that the data analysis method is in good agreement with the artificial judgment. This proposed method has a certain practical significance, which realises the data processing of dynamic threat assessment.