Bayesian Service Population Prediction Based on Sewage Flow and Weather Data Fusion
The traditional population prediction model based on the daily average sewage flow and per capita water consumption lacks consideration of weather factors,and there are problems such as overestimation of population size.In order to comprehensively consider the influence of weather factors on the daily average sewage flow,a Bayesian service population prediction model based on sewage monitoring data and weather data fusion is proposed.By introducing weather influence factors,the homogeneous and heterogeneous conversion rates of weather influence factors,and the contribution of weather factors to the daily average sewage flow,a Bayesian method based generative model is constructed.Based on stochastic variational inference,posterior distributions of the generative model parameters are obtained,and a service population prediction model for the service area of each sewage treatment plant is implemented.This model can offset the comprehensive influence level of regional weather factors and more reasonably predict the population of the sewage treatment plant service area.At the same time,statistical analysis was conducted to compare the estimation of homogeneous and heterogeneous conversion rates of weather impact factors,as well as the influence of weather factors on the daily average sewage flow rate.This service population prediction model can further support the perception of urban population trends and is of great significance for improving social governance capabilities.
sewage monitoringmulti-source data fusionservice population predictionBayesian analysisstochastic variational inference
万方数据
维普
