A Bayesian Dynamic Forecast Model Based on Neural Network

A neural network-based Bayesian dynamic forecasting model is provided in this paper. Compared with the traditional Bayesian forecasting model, the given model can also has the virtues such as it does not need the placidity suppose which is necessary in the traditional time series forecast method and it can obtain more accurate estimate even with few datum depended on the subjective priori information. In additional, the given model can also improve forecast precision for unexpected events by takes the prediction of neural network as specialistpsilas information. At last, the given model is used to forecast water supply of Shenzhen. Case study showed that the given model could enhance the forecast precision. The forecasting result is much better than that of the forecast of grey and neural network.