Application research of BP-neutral network model in forecasting sewage volume in Beijing city

In order to meet the urgent demand for sewage treatment in Beijing, three methods-PLSR, GM (1,1) and BP-Network were used to fit the sewage volume of Beijing from 2010 to 2019, and the fitting conditions of the three were compared in this paper. In this study, the application of the three methods was analyzed, and the Z-score method was used to eliminate the impact of data dimension and standardize the data. At the same time, the normal distribution of the data was tested, as was the correlation between the data, Which improved the accuracy of the three fitting models. The final result is that the realistic BP neural network has the highest fitting accuracy and can improve the most accurate sewage volume prediction for Beijing.