At present, there are not the methods of learning dynamic Bayesian network structure from no time symmetry
data. Therefore, a method of learning dynamic Bayesian network structure from no time symmetry data is developed by
combining transfer variables. In this method, first transfer variable series between two adjacent time slices are found. Then
dynamic Bayesian network part structure can be learned based on sorting nodes and local search and scoring method. Finally,
a whole dynamic Bayesian network structure can be presented by extending along time series.