Forecasting of Average Monthly Flow for Two Stations at Khabour River Using Arima And Ann Models

In this study, autoregressive integrated moving average (ARIMA) and artificial neural network (ANN) models were applied to predict the monthly flow time series of two stations, named Begova Chemcermo, for Khabour River. The analysis was performed using several criteria tests. autocorrelation function (ACF) partial (PACF) check accuracy ARIMA model. Akaike (AIC) Bayesian (BIC) equations determine optimum model prediction, which depended on lowest AIC BIC values. Applied test results show that order (0,0,0)(3,0,0)12 ARIMA((0,0,5)(5,1,4)12 have higher acceptance compared other predicting Chemecermo respectively. An ANN type multilayer perceptron method (ANN-MLP) used where best found are MLP (5,3,1) (9,7,1), Different statistical tests showed efficiency better than model, with deterministic coefficients 0.914 0.876 0.854, 0.852 respectively