Author(s): Tran Anh Duong; Nguyen Mai Dang
Linked Author(s):
Keywords: Genetic algorithm; Simulated Annealing; Rainfall Forecasting; Model Selection
Abstract: Rainfall forecasting is acritical issue on climate impact studies in hydrology due to its important role in future climate change prediction. One of the most popular techniques utilized in rainfall prediction is linear, nonlinear regression. Rainfall is complicated phenomenon, so it is more complex to simulate and predict and most challenge is to find which models are suitable for prediction. This paper attempts toapply a new method called Genetic algorithm and simulated annealing algorithm (GA-SA) to improve Autoregressive moving average (ARMA) in term of model selection and produce new flexible model; after that wecompare the SA-GA resultstoartificial neural network (ANN), Autoregressive integrated moving average (ARIMA) results by forecasting of monthly rainfall data of Ca Mau, Thanh Hoa province. The result shows that, the SA-GA model performance is better than that of ANN and ARIMA.
Year: 2014