Prediction of Water Demand in the Agricultural Sector of the Caspian Littoral Provinces: Comparison of Markov-Switching and ANN Models

In today's world, water supply and demand management plays a pivotal role in organizing and planning the drinking water supply of urban and rural residents as well as the water supply of farmers and industrialists, especially in the current situation which all countries are facing the consequences of climate change. Therefore, in this study, the water demand of the agricultural sector of the Caspian littoral provinces was predicted by Markov-Switching method and compared with the Artificial Neural Network model using seasonal data for the period 2001: 1 to 2018: 4. Comparing the efficiency of water demand models estimated by Markov-Switching and ANN methods using Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) Showed that Markov-Switching approach was more efficient than water ANN models for predicting water demand. In addition, the forecast of agricultural water demand for both seasonal and annual periods was made for the periods of 2019: 1 to 2023: 4 and 2019 to 2023, respectively.