Wavelet and neuro-fuzzy conjunction model for precipitation forecasting

Abstract

A new conjunction method (wavelet-neuro-fuzzy) for precipitation forecast is proposed in this study. The conjunction method combines two methods, discrete wavelet transform and neuro-fuzzy. The observed daily precipitations are decomposed some sub-series by using discrete wavelet transform and then appropriate sub-series are used as inputs to the neuro-fuzzy models for forecasting of daily precipitations. The daily precipitation data of three stations in Turkey are used as case studies. The wavelet-neuro-fuzzy model is provided a good fit with the observed data, especially for time series which have zero precipitation in the summer months and for the peaks in the testing period. The conjunction models are compared with classical neuro-fuzzy model. The benchmark results showed that the conjunction model produced significantly better results than the latter.