| 一种多因子加权平均温度模型的构建方法 |
| |
| 引用本文： | 李得宴,杨维芳,高志钰,蒋玉祥,李蓉蓉.一种多因子加权平均温度模型的构建方法[J].测绘科学,2021,46(3):73-79. |
| |
| 作者姓名： | 李得宴 杨维芳 高志钰 蒋玉祥 李蓉蓉 |
| |
| 作者单位： | 兰州交通大学测绘与地理信息学院,兰州 730070;地理国情监测技术应用国家地方联合工程研究中心,兰州 730070;甘肃省地理国情监测工程实验室,兰州 730070;兰州交通大学测绘与地理信息学院,兰州 730070;地理国情监测技术应用国家地方联合工程研究中心,兰州 730070;甘肃省地理国情监测工程实验室,兰州 730070;中国地震局地质研究所,北京 100029;西宁市测绘院,西宁 810000 |
| |
| 基金项目： | 科技部重大专项(2016YFB0501802);中国博士后科学基金资助项目(2019M660091XB);兰州交通大学青年科学基金资助项目(2019003);兰州交通大学优秀平台项目(201806)。 |
| |
| 摘 要： | 针对在地基GNSS水汽反演的过程中,天顶湿延迟转换为大气可降水量时如何建立精确的大气加权平均温度(Tm)模型的问题,该文在建立Tm模型前全面考虑了对Tm有显著影响的变量并选择最优回归子集。但分析发现,最优回归子集中各变量之间存在较强的相关性,这将会导致变量之间存在多重共线性,从而影响模型的稳定性和可靠性。选择2013—2015年相关气象数据作为变量并应用岭回归的方法削弱变量之间的多重共线性,建立稳定的多因子Tm回归模型。并利用该模型分别预测2016年1—12月、2019年1—7月的Tm,均方根误差分别为2.3 K和2.0 K,预测精度较高,这将为高精度的水汽反演奠定较好的数据基础。
|
| 关 键 词： | 加权平均温度 多因子线性回归 最优回归子集 多重共线性 岭回归 |
A construction method of multi factor weighted average temperature model |
| |
| LI Deyan,YANG Weifang,GAO Zhiyu,JIANG Yuxiang,LI Rongrong.A construction method of multi factor weighted average temperature model[J].Science of Surveying and Mapping,2021,46(3):73-79. |
| |
| Authors: | LI Deyan YANG Weifang GAO Zhiyu JIANG Yuxiang LI Rongrong |
| |
| Institution: | (Faculty of Geomatics,Lanzhou Jiaotong University,Lanzhou 730070,China;Nation-Local Joint Engineering Research Center of Technologies and Applications for National Geographic State Monitoring,Lanzhou 730070,China;Gansu Provincial Engineering Laboratory for National Geographic State Monitoring,Lanzhou 730070,China;Institute of Geology,China Earthquake Administration,Beijing 100029,China;Xining Surverying and Mapping Institute,Xining 810000,China) |
| |
| Abstract: | Establishing a precise atmospheric weighted mean temperature(Tm)model is a crucial part of the conversion of zenith wet delay to atmospheric precipitable water vapor(PWV)in the process of ground-based GNSS water vapor inversion.This paper comprehensively considers the variables that have a significant influence on Tm and selects the optimal regression subset before establishing the Tm model.It is found that there is a strong correlation between the variables in the optimal regression subset,which will cause the multi-collinearity between the variables,thus affecting the stability and reliability of the model.Selecting the data from 2013 to 2015 and applying the ridge regression(RR)method to weaken the multicollinearity between variables to establish a stable multi-factor Tm model.Using this model to predict the Tm between 2016 and July 2019.The root mean square error is 2.297 K and 2.02 K,respectively.The prediction accuracy is high,which will lay the foundation for high quality and high precision water vapor inversion. |
| |
| Keywords: | weighted mean temperature multi-factor linear regression optimal regression subset multi-collinearity ridge regression |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
