A New Fusion Model for Precipitation Prediction

In this paper, a new fusion model is established based on the regression (ARIMA) model and BP (Back Propagation) neural network prediction model, which is used to predict the precipitation in Henan Province. In the Henan Statistical Yearbook, the precipitation in Henan can be obtained for nearly 20 years. Firstly, the differential integrated mobile mean auto regression (ARIMA) model is used to fit the future data according to the relationship between the precipitation and the time series. Secondly, BP (back propagation) neural network prediction model is used to establish a linear relationship between dependent and independent variables. These two models have both their own disadvantages, so we can combine the two models together to get a new fusion model. The precipitation of Henan province in the next 8 years can be well predicted by using the new fusion model.