《Table 3 Diagnostics for spatial dependence》

  提示：宽带有限、当前游客访问压缩模式

本系列图表出处文件名：随高清版一同展现

《Spatio-temporal Characteristics and Geographical Determinants of Air Quality in Cities at the Prefecture Level and Above in China》

1. 获取 高清版本忘记账户？点击这里登录

1. 下载图表忘记账户？点击这里登录

Model 1 in Table 2 is a general linear regression model，and the least squares method is used for estimates.Model 1 shows that the regression equation model is significant，with an F value of 24.56，and the significant level of 1 is tested.The coefficients from the OLS model indicate that only the effect of GDP per capita is not significant，and the resident population and secondary proportion have positive influences on the urban AQI of China.The urbanization level，DEM，NDVI，air temperature，and wind speed explanatory variables have obvious mitigating effects on urban AQI.However，because the general linear regression model does not consider the spatial correlation of urban air quality，the estimation results from the model may be biased.Therefore，the spatial correlation of the residual error of the ordinary linear regression model must be evaluated.Table 3 shows that the residual error of the general linear retrospective model is spatially correlated，and the robust LM（lag）is more significant than the robust LM（error）；therefore，the spatial hysteresis model is more appropriate.