Black box models are one of the most common hydrological
models in order to make predictions of hydrological variables such as
precipitation and stream flow. In this study, performance of a combined model
which consists of wavelet transformation, genetic algorithm and artificial
neural network (WGANN) were tested for
prediction of monthly precipitation by using North Atlantic Oscillation (NAO)
index, Southern Oscillation (SO) index and precipitation data as input in the
model. The case study was carried out for Antalya which is located in Mediterranean
region of Turkey. As a result, it was attained that WGANN model performed more
successful than usual artificial neural network (ANN), multiple linear
regression (MLR) and genetic algorithm-artificial neural network (GANN) models.
Artificial neural network genetic algorithm precipitation regression wavelet transformation
Konular | Mühendislik |
---|---|
Bölüm | Makaleler |
Yazarlar | |
Yayımlanma Tarihi | 9 Kasım 2017 |
Yayımlandığı Sayı | Yıl 2017Sayı: 1 |