The research of CREAM prediction analysis method based on BP neural network under dynamic context

Although the basic method of cognitive reliability and error analysis method (CREAM) is widely used, there are still a lot of problems, for example, there is no consideration of the problems that CPC has different weights in different industrial environments and the process of determining control mode is not smooth. Therefore, the prediction of human error probability (HEP) in the basic method is not accurate. For this reason, the method of predicting HEP based on BP neural network is proposed. Firstly, the context is quantified by the way of expert scoring. Then, based on related data, the function relationship between the HEP and the context is fitted by utilizing strong nonlinear data fitting ability of BP neural network. Finally, the HEP can be predicted by the gotten fitting function accurately.