C.L. Wu، نويسنده , , K.W. Chau and J.S. Huang، نويسنده , , C. Fan، نويسنده ,
This study is an attempt to seek a relatively optimal data-driven model for rainfall forecasting from three aspects: model inputs, modeling methods, and data-preprocessing techniques. Four rain data records from different regions, namely two monthly and two daily series, are examined. A comparison of seven input techniques, either linear or nonlinear, indicates that linear correlation analysis (LCA) is capable of identifying model inputs reasonably. A proposed model, modular artificial neural network (MANN), is compared with three benchmark models, viz. artificial neural network (ANN), K-nearest-neighbors (K-NN), and linear regression (LR). Prediction is performed in the context of two modes including normal mode (viz., without data preprocessing) and data preprocessing mode. Results from the normal mode indicate that MANN performs the best among all four models, but the advantage of MANN over ANN is not significant in monthly rainfall series forecasting. Under the data preprocessing mode, each of LR, K-NN and ANN is respectively coupled with three data-preprocessing techniques including moving average (MA), principal component analysis (PCA), and singular spectrum analysis (SSA). Results indicate that the improvement of model performance generated by SSA is considerable whereas those of MA or PCA are slight. Moreover, when MANN is coupled with SSA, results show that advantages of MANN over other models are quite noticeable, particularly for daily rainfall forecasting. Therefore, the proposed optimal rainfall forecasting model can be derived from MANN coupled with SSA.
Rainfall prediction , Modular artificial neural network , Principal component analysis , Moving average , Fuzzy c-means clustering , Singular spectral analysis