Urmia Lake level forecasting using Brain Emotional Learning (BEL)

This paper has tried to focus on a new approach for water level forecasting of Urmia Lake by using records of past time series and emotional learning. Water level forecasting is important in water resources engineering and management and efficient management of water resources for use. During the past two decades, the approaches artificial intelligence based on the Genetic Programming (GP), Artificial Neural Networks (ANN), fuzzy logic, neuro-fuzzy and statistical method for example ARIMA and recently, chaos theory have been developed. Time series the measurements from tide gauge at Urmia Lake, were used to train emotional learning approach for the period from March 1965 to February 2011. The research indicates that there is a non-linear and complex relationship between water input and variables, therefore anticipation seems to be more difficult to implement it with conventional tools of time series prediction. Simulation results prove that the applied method has prominent capability in forecasting time series. In this paper, various criterion including Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE) have been used.