在模型比较中,有很多评价标准,如p-值等,都受制于数据的分布假定。而利用交叉验证法进行数据处理,然后比较归一化均方误差Normalized Mean Squared Error (NMSE)是目前最流行的模型评价的标准,不受任何数据分布的限制。本文详细介绍了...在模型比较中,有很多评价标准,如p-值等,都受制于数据的分布假定。而利用交叉验证法进行数据处理,然后比较归一化均方误差Normalized Mean Squared Error (NMSE)是目前最流行的模型评价的标准,不受任何数据分布的限制。本文详细介绍了交叉验证法,并给出了其具体的应用。通过对实际的问题建立了6种不同的模型,并利用10折交叉验证法对不同模型的归一化均方误差(NMSE)进行比较,选择出了最优的预测精度最高的模型。展开更多
This paper presents a rapid regression algorithm for the retrieval of methane(CH4)profile from Atmospheric Infrared Sounder(AIRS)based on empirical orthogonal functions(EOF)and its validation.This algorithm was traine...This paper presents a rapid regression algorithm for the retrieval of methane(CH4)profile from Atmospheric Infrared Sounder(AIRS)based on empirical orthogonal functions(EOF)and its validation.This algorithm was trained using the simulated radiance from an assemble of atmospheric profiles and can be utilized to derive the CH4profile rapidly with the input of the AIRS cloud-clear radiance.Validation using hundreds of aircraft profiles demonstrates that the root mean square error(RMSE)is about 1.5%in the AIRS sensitive region of359–596 hPa,which is smaller than AIRS-V5 product(except in high latitudes).Comparison with the groundbased solar Fourier transform spectrometry observations showed that the RMSE of the retrieved CH4total column amount is less than 3%.This EOF-based regression method can be easily applied to other thermal infrared sounders for deriving CH4and some other gases,and the derived profiles can be used as the first guess for further physical retrieval.展开更多
文摘在模型比较中,有很多评价标准,如p-值等,都受制于数据的分布假定。而利用交叉验证法进行数据处理,然后比较归一化均方误差Normalized Mean Squared Error (NMSE)是目前最流行的模型评价的标准,不受任何数据分布的限制。本文详细介绍了交叉验证法,并给出了其具体的应用。通过对实际的问题建立了6种不同的模型,并利用10折交叉验证法对不同模型的归一化均方误差(NMSE)进行比较,选择出了最优的预测精度最高的模型。
基金supported by the Strategic Priority Research Program – Climate Change: Carbon Budget and Relevant Issues of the Chinese Academy of Sciences (XDA05090101)the Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences (Y1S02000CX)
文摘This paper presents a rapid regression algorithm for the retrieval of methane(CH4)profile from Atmospheric Infrared Sounder(AIRS)based on empirical orthogonal functions(EOF)and its validation.This algorithm was trained using the simulated radiance from an assemble of atmospheric profiles and can be utilized to derive the CH4profile rapidly with the input of the AIRS cloud-clear radiance.Validation using hundreds of aircraft profiles demonstrates that the root mean square error(RMSE)is about 1.5%in the AIRS sensitive region of359–596 hPa,which is smaller than AIRS-V5 product(except in high latitudes).Comparison with the groundbased solar Fourier transform spectrometry observations showed that the RMSE of the retrieved CH4total column amount is less than 3%.This EOF-based regression method can be easily applied to other thermal infrared sounders for deriving CH4and some other gases,and the derived profiles can be used as the first guess for further physical retrieval.
