Research On Air Quality Prediction And Selection Of Atmospheric Environmental Governance Policy Tools ——Evidence From Beijing City

Urban development should take into account both "hard power" and "soft power".A beautiful environment is the key to maintaining the value of the city.The unsatisfactory air quality will cause the city’s commercial value to gradually lose,and the city’s prosperity will "go downhill" accordingly.Since the "Eleventh Five-Year Plan",the State has established the Ministry of Environmental Protection and issued relevant policies to control air pollution.Carry out real-time air pollution monitoring and air quality prediction in many places,and the environmental protection department will implement a series of policy measures to strive for scientific prevention and control,respond flexibly,and win the blue sky defense battle.The 20th Winter Olympics in 2022 will be jointly organized by Beijing and Zhangjiakou.During the Winter Olympics,friends from all over the world will gather in Beijing.While providing caring services,creating high-quality air quality is also one of the priorities of the work.As a result,Beijing’s air quality improvement work should still be accelerated.However,focusing on actual work,the data collected by the air quality monitoring station is complex,and the establishment of a scientific prediction model can improve the prediction accuracy,and also provide strong data and theoretical support for follow-up actions;however,reliable data prediction is not enough It also needs to be combined with effective governance measures.Only by understanding the implementation status of various policy tools can relevant departments achieve the goals of environmental governance policies faster,more accurately and more accurately.In this paper,a series of related statistical analysis is carried out on the PM2.5 concentration data in Beijing.First,analyze the PM2.5 concentration in Beijing from March 2015 to February 2019 and the hourly PM2.5 concentration data from March 2018 to February 2019.The study found that Beijing’s atmospheric environmental quality problems are more severe in winter.In the winter of 2018-2019,compared with the same period of last year,the number of severely polluted weather increased by 11 days.The average concentration of PM2.5 in Beijing in 2019 still exceeds national standards(35（Micrograms/cubic meter）20%,the peak daily PM2.5 concentration in winter reaches as much as 5.2 times the national standard,and air quality problems in Beijing are still relatively serious at this stage.Secondly,carry out prediction research on Beijing PM2.5 data.Using EEMD decomposition technology and normalization principle to perform data decomposition and noise reduction on the data sequence,combined with the SVR model optimized based on artificial intelligence algorithm,the seasonal multi-step prediction of PM2.5 concentration in Beijing.Finally found that:in terms of data preprocessing,the prediction model established by the data after EEMD decomposition has better model results than the single SVR model and EMD-SVR model prediction results;in terms of intelligent optimization algorithms,the prediction accuracy of the optimized SVR model is higher than the SVR model Prediction accuracy.Based on the CS algorithm,the article optimizes it.The improved EEMD-SDCS-SVR model predicts better than the EEMD-CS-SVR model.The former improves the model prediction accuracy by about 0.8%compared to the latter;In the step prediction section,the article develops multi-step predictions for different seasons on PM2.5 data.The research results show that the accuracy of the prediction results gradually decreases as the number of prediction steps increases.Comparing the prediction results of different data sets,the accuracy of the single-step prediction results It is generally the highest,followed by two steps,and the third step is the worst.Finally,after modeling and predicting air quality,the article uses the average annual PM2.5 concentration in Beijing as an interpreted variable to conduct an environmental policy tool selection study.A hybrid model（GRA-VAR model）combining gray correlation model and vector autoregressive model is adopted.Taking the average annual PM2.5 concentration in Beijing from 2001 to 2017 as the explanatory variable,the environmental governance policy tools issued during the period are divided into There are three categories of policy control,economic incentive and public participation policy tools.After analyzing the implementation effects of environmental governance policy tools,it is found that the sewage charge and the annual average concentration of PM2.5 are inversely related.The three parts of industrial structure and foreign investment are analyzed.The main innovations of this article are as follows:First,the air quality of Beijing in the winter of 2015 to 2018 is compared,and the daily and sub-seasonal concentrations of PM2.5 in Beijing are multi-step forecasted.In EEMD and SVR Based on the combination of the models,the cuckoo algorithm is used to optimize the model,and the EEMD-SDCS-SVR model is established to study and discuss the applicability of the model under different data sets.Second,select the gray correlation degree and the vector autoregressive model（GRA-VAR）to analyze the atmospheric environmental governance policy tools of specific cities in Beijing.The GRA-VAR model can reflect the internal relations between variables,and the hybrid model can also compensate for the gray correlation analysis.The model cannot judge the lack of positive and negative correlations;meanwhile,it analyzes the relationship between sewage charges and the average annual concentration of PM2.5.