Annual Rainfall Prediction Using Neuro-Fuzzy and Wavelet Approach

Rainfall prediction has become an integral part of the hydrological model. This paper presents the method of annual rainfall forecasting using wavelet transform and neuro-fuzzy approach based on the past history. The historical data have been decomposed into wavelet domain constitutive sub-series. The behaviour of the wavelet domain constitutive series has been studied based on the statistical analysis. Forecasting performance of the wavelet-coupled model has been compared with classical neuro-fuzzy, neural network and multiple linear regression models. The benchmark result shows that wavelet coupled model produces significantly better results in comparison with the neuro-fuzzy, neural network and regression models.