A Combined Model from Artificial Neutral Network, Wavelet Converting, and ARMA in Predicting the Demand for Urban Water
Predicting the demand for urban water helps the managers and users of urban water systems to a great extent in order to manage properly in this regard. To do so، exact predicting of the demand for water is of great importance in different periods. In this study، by using a method which is combination of linear and non-linear models، daily demand for water in Tehran and the effective factors on daily demand has been studied. Daily demand for urban water is predicted step by step based on ARMA models، Artificial Neutral Network، and the combined model for the next 10 days. To design Artificial Neutral Network and the combined model، the effective factors on daily demand for urban water، temperature (minimum، maximum، and average (medium))، weekly days، holidays، and the special days are considered. The results obtained from using criterion for accuracy evaluation show that the combined model has fewer error and high accuracy for this propose and Artificial Neutral Network and ARMA process are in next priorities.