Research On Forecasting And Modeling Of Urban Daily Water Consumption Data Based On Evidence Theory

With the continuous development of cities,more and more problems arise from the traditional methods of water resources management in the integrated decision-making of water resources.Accurate prediction of urban water consumption is the premise of supporting the comprehensive decision-making and management of urban water consumption.In recent years,many domestic and foreign scholars have done a lot of research on water consumption prediction,mainly from three aspects:First,from the water consumption influence factors,the establishment of a mathematical model about the water consumption influence factors to predict;Secondly,based on the historical data,the data mining and data analysis method are used to build a mathematical model according to the sample data and to predict by empirical data.Third,from the perspective of probability,the uncertainty reasoning method mainly includes Bayes method,possibility theory,determinism theory,evidence theory and a series of reasoning models.There is no one-size-fits-all method for the prediction of urban water resources.In this paper,the existing urban daily water consumption data prediction algorithms are analyzed and deduced theoretically.from the two aspects of the universal applicability of water consumption prediction models and the real-time performance of the algorithms,it is found that the evidence theory has unique advantages in the urban water consumption prediction.Based on the conflict coefficient of evidence theory,a combination rule of evidence theory based on the global conflict coefficient is proposed,which is improved by combining the original combination rule of evidence theory.Then the improved combination rule of evidence theory is applied to the prediction of urban daily water consumption.The main research work of this paper is as follows:1.The evidence theory is chosen as the prediction model of urban water use to study the source of paradox and to find another factor that causes the paradox of evidence theory.By comparing the algorithms of prediction models of urban water consumption,we find that: the number of samples needed by evidence theory in various prediction models is the smallest,and the evidence theory has excellent mathematical characteristics and perfect uncertainty reasoning theory in uncertaintyreasoning.The original evidence theory combination conflict evidence is contrary to common sense.Based on the existing factors of the combination rule of evidence theory,this paper proves that the paradox mainly comes from the definition of conflict coefficient by demonstration.The idea of defining the conflict coefficient from the conflict between the evidence and the conflict between the focal elements is put forward.2.To improve the existing combination rules of evidence theory and to propose a combination rules of evidence theory based on global conflict coefficient.The improved combination rule redefines the conflict coefficient and deduces the combination rule of evidence theory based on global conflict.The improved combination rules are contrary to common sense when the original combination rules are in conflict with each other.The expression of conflict coefficient is improved.Based on the characteristic of the basic probability distribution function,the new combination rules are deduced from the new evidence theory.By comparing the combination results of normal evidence and conflicting evidence with other improved algorithms,it is proved that the improved algorithm has better stability and credibility accumulation ability in dealing with conflicting evidence,and has better convergence characteristics in normal evidence combination.3.The urban daily water consumption prediction model based on evidence theory is established and compared with other prediction models.The daily water consumption time series update the sample data iteratively before the prediction,and establish the dynamic prediction model of evidence theory data fusion algorithm.Daily Water Consumption Prediction Using Improved Evidence Theory Data Fusion Model and Support Vector Machine Prediction Model,The comparison between the prediction results and the actual water consumption shows that: the evidence theory data fusion algorithm needs few samples,is low affected by the historical data,has better prediction accuracy,and is more suitable for the prediction of the daily water consumption of the city.