Analysis And Forecasting Of Lanzhou Air Pollution Index Based On Data Preprocessing And Machine Learning Hybrid Model

The atmosphere is one of the indispensable environmental factors for human survival and development.However,with the rapid development of modern industry and the rapid growth of urban population,a large number of fossil fuels are burning.The chemical substances produced by combustion are discharged into the atmosphere in the form of exhaust gas and smoke,which exceed the allowable amount of atmospheric environment,and bring serious impact on human life,production and health.A report published by the WHO and the UN Environment Organization said: "air pollution has become an inescapable reality in the lives of urban residents around the world." Therefore,the study of air pollution prediction can prevent the occurrence of serious pollution events,help government management and decision-making departments to take measures to the potential events,and provide the reference basis for the public social activities.In this paper,a new hybrid prediction model based on BP neural network is proposed.In order to evaluate the accuracy and effectiveness of the prediction model,the historical data of the atmospheric particles PM2.5 and PM10 in 2016,Lanzhou,Gansu,China were selected as a case to be analyzed and predicted.In order to establish an effective prediction model for PM2.5 and PM10,firstly,the Variational Mode Decomposition(VMD)technology is used to reconstruct the data in the data preprocessing stage.Secondly,an improved ant colony based on Adaptive Particle Swarm Optimization(APSO)is proposed to optimize the weights and thresholds of the BP neural network.Algorithm(ACO)hybrid optimization algorithm.Finally,the prediction results are compared and analyzed in various ways.The experimental results show that the hybrid prediction model(VMDAPSOACO-BP)proposed in this paper has higher prediction accuracy.At the same time,the model basically fits the trend and fluctuation of the atmospheric particle concentration change,so it can provide a certain decision support for the local government to predict and study the pollution index.