建立变系数部分线性固定效应模型分析PM2.5与其他空气污染物的相关关系

以陕西省西安市13个空气观测地点的污染物观测资料为依据，运用统计学分析方法，建立了研究大气污染物PM2.5的浓度与包括当天和前一天SO2、NO2、PM10、CO、O3在内的11个观测指标在空气中浓度相关关系的模型。首先通过做散点图和计算相关系数来确定PM2.5和所选解释变量间相关关系，继而根据结果建立变系数部分线性固定效应模型，同时估计模型中各参数，并对模型中线性部分做变量选择，从而近一步优化所建立的模型。最后，将改进后模型的计算结果与实际情形对比分析，证实本文所得结论与实际情形相符合，表明所建模型良好。

A varying coefficient partially linear fixed effects model for analysis of the correlation between PM2.5 and other air pollutants

The correlation between PM2.5concentrations and the concentrations of SO2, NO2, PM10, CO, O3 on the same day and the previous day has been studied using a statistical analysis method. The experimental data were obtained from thirteen air pollutant observation points in Xi'an, Shaanxi Province. Firstly, the relationship between PM2.5 and selected explanatory variables was determined by means of a scatter plot and by calculating the correlation coefficients. Then, according to the calculation results, a semi-parameter varying coefficient fixed effects model was set up. We combined the profile least squares method and the local linear smoothing method to estimate the parameters in the model, and then used the SCAD method to select the variables for the linear section of the model. Finally, the results of the improved model and the actual situation were compared, and the calculation results were consistent with the actual situation, which showed the accuracy of the improved model.