Study on Combination of ANN Method with Mean Generating Functions for Short-Term Climate Prediction

摘要

Artificial neural network (ANN) and mean value generating functions in combination are employed to construct a nonlinear long-range prediction model to investigate climate variation on an annual basis for 2 ～ 3 years to come with more success and another important feature of the model is that it can well explain correct and wrong results for they have greatly differing structures of output both from fitting and prediction. The study may open up a new way to this branch of science.