Prediction of wastewater discharge based on GRA-LSTM: a case study of Beijing

Water resources, as one of the indispensable resources for urban development, have become an important factor limiting the sustainable development of cities. In order to promote sustainable urban development, Beijing has set the work task of reaching 99% of urban and rural wastewater treatment rate from 2020 to 2035. Accurate prediction of future wastewater discharge is essential to achieve the target. For this reason, this study takes Beijing as the research object and constructs a combined prediction model based on gray relational analysis and long- and short-term memory (GRA-LSTM). Firstly, gray relational analysis (GRA) is used to analyze the correlation of the experimental data indicators affecting the amount of wastewater discharged in order to obtain experimental data indicators with stronger correlation. Secondly, the long short-term memory (LSTM) model was used to learn the characteristics of the key impact indicators and obtain the optimal model parameters. The results showed that the mean absolute percentage error (MAPE) value of the combined GRA-LSTM model constructed in this study was 5.62%, and the prediction accuracy was higher than that of the other seven prediction models. Then, three scenarios with low, medium, and high dimensions were set to predict the wastewater discharge in Beijing from 2020 to 2035, and the prediction result that the wastewater discharge in Beijing will still continue to grow was obtained. Finally, in order to improve the water utilization rate and promote the sustainable development of the city, this study proposes relevant policy recommendations in terms of the unbalanced urban-rural development of Beijing's wastewater treatment capacity and the increase of recycled water usage.