Water Demand Prediction in Haining City Based on Improved PCA-BP Neural Network Model
Water demand prediction is an important part of regional water resources planning,which is of great sig-nificance for realizing the rational and orderly development of water resources and guaranteeing the sustainable develop-ment of society and economy.The article adopts the improved PCA-BP neural network model to reduce the dimensionali-ty of the nine influencing factors affecting water demand,and completes the training of the model with the data of Haining City in 2001-2014 and 2015-2020 as the training samples and test samples,respectively,in which the integrated grey pre-diction model GM(1,1)independently predicts the influencing factors after the reduction of dimensionality so as to predict the annual water demand of Haining City in the planning year.Finally,the traditional quota method is used to predict the annual water demand of the planning year and its comparative analysis.The results show that the population,GDP,resi-dential water consumption,urban public water consumption are the main factors affecting the water demand of Haining City;The water demands in 2025,2030 and 2035 obtained by the improved PCA-BP neural network model are more real-istic and reasonable than the traditional quota method,which further confirms the reasonableness of the prediction model and can provide the corresponding guidance for the future water resources planning of Haining City.
water demand predictionprincipal component analysisimproved PCA-BP neural networkgrey predic-tion model
NETL
NSTL
万方数据
