Application of Fuzzy Neural Network to the Flood Season Precipitation Forecast

摘要

Taking the flood season (from May to October) precipitation in Hainan province as the forecast object, the application of fuzzy neural networks forecasting method with different forecast factors is studied. The results show that the new model based on principal component analysis is significantly superior to the traditional stepwise regression model and other fuzzy neural networks models which select other factors in prediction accuracy and prediction stability. It can be applied to operational short-term climate forecast.