Rainfall time series forecasting based on Modular RBF Neural Network model coupled with SSA and PLS

Accurate forecast of rainfall has been one of the most important issues in hydrological research. Due to rainfall forecasting involves a rather complex nonlinear data pattern; there are lots of novel forecasting approaches to improve the forecasting accuracy. In this paper, a new approach using the Modular Radial Basis Function Neural Network (M–RBF–NN) technique is presented to improve rainfall forecasting performance coupled with appropriate data–preprocessing techniques by Singular Spectrum Analysis (SSA) and Partial Least Square (PLS) regression. In the process of modular modeling, SSA is applied for the time series extraction of complex trends and structure finding. In the second stage, the data set is divided into different training sets by Bagging and Boosting technology. In the third stage, the modular RBF–NN predictors are produced by a different kernel function. In the fourth stage, PLS technology is used to choose the appropriate number of neural network ensemble members. In the final stage, least squares support vector regression is used for ensemble of the M–RBF–NN to prediction purpose. The developed RBF–NN model is being applied for real time rainfall forecasting and flood management in Liuzhou, Guangxi. Aimed at providing forecasts in a near real time schedule, different network types were tested with the same input information. Additionally, forecasts by M–RBF–NN model were compared to the convenient approach. Results show that the predictions made using the M–RBF–NN approach are consistently better than those obtained using the other method presented in this study in terms of the same measurements. Sensitivity analysis indicated that the proposed M-RBF-NN technique provides a promising alternative to rainfall prediction.