Prediction Of The Water Consumption In Yangquan City Based On The RBF Neural Network Model With DPSIR-principal Component Analysis

Yangquan City is one of the most water-scarce areas in Shanxi Province.The distribution of water resources in Yangquan City is uneven in time and space,the problem of water pollution is serious,and the limited water resources are difficult to be developed and utilized sustainably.In order to realize the sustainable development and utilization of water resources and the high quality development of economy,it is necessary to plan,allocate and schedule water resources reasonably.Accurate prediction of water consumption is the premise of solving problems related to water resources.According to the basic water use situation of Yangquan City,the DPSIR model of water resources in Yangquan City is established,and the index factors affecting the water consumption and sustainable utilization of water resources in Yangquan City are classified according to the driving force,pressure,state and influence.Screening and analysis.Among the five criterion layers,35 indexes that can be quantitatively calculated are selected for driving force,pressure,state and influence,and onequalitative index is selected for reference analysis in the response criterion layer.Based on the data of water consumption indexes collected from 2005 to 2017 in Yangquan City,principal component analysis(PCA)was used to analyze the water consumption of Yangquan City.The index of each standard layer is weighted by dimension reduction,and the principal component index is formed,and it is defined as a reasonable prediction index.The prediction index is saved as the input factor of RBF neural network model,and the rationality analysis of water consumption prediction in Yangquan City is realized by using neural network tool model newrb.The prediction results are compared with the model of RBF neural network system which only carries out principal component analysis to predict water consumption in Yangquan City.Compared with the total water consumption values of Yangquan City in2015,2016 and 2017,the prediction results of RBF neural network model based on DPSIR-principal component analysis have a smaller error,with an average relative error of 0.42%.The error between the predicted value of the model and the total water consumption value of Yangquan City from 2015 to 2017 is 9.6%.The index selection of RBF neural network model based on DPSIR-principal component analysis is comprehensive and representative,which can accurately reflect the factors that affect the prediction of water consumption in Yangquan City.The evaluation index is at the end of the evaluation index.In the process of continuous operation,the prediction process is more simple,and the results are more accurate than other models.It is a more reliable method for predicting thefuture water consumption of small and medium-sized cities.