Comparison of Artificial Intelligence Methods in Predicting Daily Time Series of Minimum and Maximum Temperature and Precipitation in Tangab Dam Station (Fars Province)

Development of a prediction hydrological model based on past records depends on the proper prediction and understating of time series effective on water resources to manage and plan water reservoirs effectively. In recent years, a growing issue in this context is the application of artificial intelligence techniques in modeling, forecasting and recovery of hydrological data. This paper compares the artificial intelligence methods in predicting and recovery of time series of daily minimum and maximum temperatures and precipitation in Tangab dam station. Both series (using delay in the series) and nearby stations are used in this study to recover and predict data. Multi-layer perceptron (MLP), radial basis functions(RBF), support vector machine (SVM), fuzzy inference system (FIS) and adaptive neuro-fuzzy inference system (ANFIS) methods have been studied. In order to evaluate the performance of these models, the mean squared error (MSE), correlation coefficient (R), variance and standard deviation of obtained data, as well as graphical diagrams have been used. The results showed the inefficiency of the models in predicting precipitation, but these can be used in recovering the precipitation and predicting temperature.