|
| 基于PCA-GWR模型的PM2.5浓度空间插值 |
| Spatial Interpolation of PM2.5 Concentration Based on PCA GWR Model |
| |
| DOI： |
| 中文关键词: PM2.5 地理加权回归 主成分分析 径向基函数插值 |
| 英文关键词: PM2.5 Geographic weighted regression Principal component analysis Radial basis function interpolation |
| 基金项目:国家自然科学基金资助项目（41864002） |
|
| 摘要点击次数: 924 |
| 全文下载次数: 134 |
| 中文摘要: |
| 采用主成分分析（PCA）和地理加权回归（GWR）模型，结合大气污染物、气象和气溶胶光学厚度（AOD）等数据，估算长三角地区的PM25浓度。引入两种径向基函数（高次曲面和反高次曲面）插值方法对PCA-GWR模型残差进行改正，构建两种基于残差改正的PCA-GWR模型，以优化模型精度。结果表明，PCA-GWR模型能够充分考虑原始变量的信息，有效降低解释变量之间的多重共线性；基于两种插值法改正的PCA-GWR模型的R2、RMSE和MAE相较PCA-GWR模型均有明显提升，其中基于高次曲面插值法（PCA-GWR-MQ）的模型具有更好的插值精度，在不同时间尺度的稳定性更佳。 |
| 英文摘要: |
| Principal component analysis（PCA） and geographically weighted regression（GWR） models were used to estimate PM25 concentration in the Yangtze River Delta region，according to the data on atmospheric pollutants，meteorology and aerosol optical thickness（AOD）. Two radial basis function（high order surface and anti high order surface） interpolation methods were introduced to correct the residuals of the PCA GWR model， and two PCA GWR models based on residual correction were constructed to optimized the model accuracy. The results showed that the PCA GWR model could fully consider the information of the original variables and effectively reduce the multicollinearity among the explanatory variables. R2，RMSE and MAE of the PCA GWR model corrected by the two interpolation methods were significantly improved compared with those of the original model. Among them， the model based on high order surface interpolation（PCA GWR MQ） had higher interpolation accuracy and superior stability at different time scales. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
