Predicting the drinking water consumption by equipping hybrid artificial neural network with discrete wavelet transform

In this research, to predict the amount of water consumption (demand), the artificial neural network (ANN) equipped with discrete wavelet transform (DWT) function and a new model named W-ANN is proposed. In addition, Pierson correlation coefficient is applied for finding the proper input data set. Here, data set includes temperature, precipitation, humidity and the amount of daily water produced from the beginning of ۲۰۱۴ to the end of ۲۰۱۹. To evaluate the performance of the models, statistical indices including R², RMSE and NSE have been calculated. Comparison of the results showed that using discrete wavelet transform function to proposed W-ANN model improves the performance of the proposed models and their results. In addition, best results are obtained by using W-ANN model. In which for training data, the RMSE, NSE, R۲ values are ۲۸۷۷.۳۷ (MCM), ۰.۸۳ and ۰.۸۶, and for test data are ۳۴۷۰.۸۱ (MCM), ۰.۷۴ and ۰.۷۶.