Improving Rainfall Forecasting via Radial Basis Function and Deep Convolutional Neural Networks Integration

Abstract

The foremost challenge of rainfall forecasting is the intensity of rainfall in some particular stations. The unpredictable rainfall volume owing to the climate transformation can root cause for either overflow or dryness in the reservoir. In this article, we coin a novel model to predict the monthly rainfall by using an Ensemble Radial basis function Network and a One-Dimensional Deep Convolutional Neural Network algorithm. In the first step, nine climatological parameters, which are highly related to monthly rainfall disparity, are given as input for an ensemble model. In the second step, a hybrid approach is proposed and compared with Bayesian Linear Regression (BLR) and Decision Forest Regression (DFR). Experimental results show that the ensemble approach yields good results in seizing the multifaceted association among causal variables and also it extracted the most relevant hidden features of hydro meteorological rainfall system.