Local Linear Kernel Estimation Of Nonparametric Spatial Lag Model And Its Application To Air Quality Evaluation In Shandong Province

With the development of society,people begin to pursue a more comfortable living environment after meeting their basic physiological needs,so the issue of air quality has been paid more and more attention by people.At the same time,Shandong Province,as one of the traditional industrial provinces,leads to the high level of air pollutant emissions.Therefore,in order to understand the overall air quality changes in Shandong Province,the following studies were conducted in this paper:In this paper,the data of PM_2.5,O₃and meteorological variables（air temperature,air pressure,wind speed,dew point temperature and precipitation）every 3 hours of 16cities in Shandong Province from August 2018 to April 2022 are used,based on the non-parametric spatial lag model,and the city’s virtual variables are introduced into the model,and the local linear kernel estimation method is used to study the change trend of different pollutants.Then the overall air quality in Shandong Province was evaluated effectively.The advantages of this method are as follows:（1）Considering the effect of spatial effect on the overall air quality in Shandong Province;（2）Compared with the parameter estimation method,the model estimation is more flexible,and the model estimation is completely driven by data,which can effectively avoid the influence of subjective experience on the model estimation;（3）Compared with nonparametric kernel density estimation,boundary effect can be avoided effectively.The data are then tested using Fisher’s test and Moran’s index,and correlations between variables are analyzed.The experimental results show that:（1）The values of the four statistics of Fisher’s test are far less than 0.05,that is,PM_2.5,O₃and 5meteorological factors are all sequentially stable.（2）In Shandong Province,the Moreland indices of PM_2.5and O₃are both positive,and the corresponding values are much less than 5%,that is,there is a significant spatial positive correlation,but the spatial difference of PM_2.5and O₃in each city has no significant change.（3）Correlation analysis of variables shows that there is interaction between meteorological variables,so it is very necessary to use non-parametric method.Finally,the empirical analysis of PM_2.5and O₃is carried out respectively.The analysis results show that:（1）The goodness of fit of PM_2.5and O₃model test is mostly above 60%,which indicates that the modeling of data by non-parametric model is reasonable.（2）The calculated monthly average PM_2.5concentration in Shandong Province showed significant seasonal variation characteristics,and showed a downward trend.Although PM_2.5decreased significantly during the lockdown period,it did not decrease significantly during other periods.（3）The calculated trend of monthly mean O₃concentration in Shandong Province also has a significant seasonal variation,but the overall trend change is not obvious.Although the concentration of O₃in May to July 2019 is higher than that in the same period of the year,this is mainly due to high temperature,and the fluctuation range of other periods is not large.The above analysis results indicate that the control of PM_2.5in Shandong Province has achieved certain results,but more efforts are needed to control O₃.