一种基于主成分分析的协同克里金插值方法

发布时间：2018-04-01 01:19

本文选题：主成分分析　切入点：协同克里金插值　出处：《测绘通报》2017年11期

【摘要】：针对协同克里金插值方法在插值时,辅助变量较多造成计算复杂度增加,而辅助变量较少引起插值精度降低这一问题,提出了一种基于主成分分析的协同克里金插值方法(PCA-CoKriging)。该方法首先使用主成分分析对插值相关变量进行将维,得到较少几个综合指标,然后里利用这几个综合指标作为辅助变量进行协同克里金插值。为验证该方法的有效性和数据分布对该方法的影响,本文选取了2016年北京市范围内4个季节中PM2.5浓度满足正态分布效果不同的4组数据,分别使用PCA-CoKriging和普通克里金插值方法、常规协同克里金插值方法,进行了插值试验。结果表明,本文方法与普通克里金插值方法、常规协同克里金插值法在4组试验中的平均绝对误差分别为4.91、6.04、5.61,平均均方根误差分别为6.65、8.76、7.57。综合比较,本文方法比常规协同克里金插值的平均绝对误差与均方根误差分别提升了10.73%、12.56%,比普通克里金插值法的平均绝对误差与均方根误差分别提升了18.71%、24.09%。
[Abstract]:In order to solve the problem that there are more auxiliary variables and less auxiliary variables to reduce the interpolation precision, a cooperative Kriging interpolation method based on principal component analysis (PCA) is proposed in this paper.In this method, we first use principal component analysis (PCA) to dimension the relevant variables of interpolation, and obtain a few synthetic indexes, and then use them as auxiliary variables to carry out cooperative Kriging interpolation.In order to verify the validity of the method and the effect of data distribution on the method, four groups of data with different normal distribution effects in four seasons in Beijing in 2016 are selected in this paper. PCA-CoKriging and Kriging interpolation methods are used respectively.The interpolation experiment was carried out by the conventional cooperative Kriging interpolation method.The results show that the average absolute error and mean root mean square error of the method are 4.91 ~ 6.04 ~ 5.61 and 6.65 ~ 8.7 ~ 7.57, respectively, compared with the ordinary Kriging interpolation method and the conventional synergetic Kriging interpolation method in the four groups of experiments.In a comprehensive comparison, the average absolute error and root mean square error of this method are increased by 10.73 and 12.56, respectively, and 18.71 and 24.09, respectively, compared with the average absolute error and root mean square error of the ordinary Kriging interpolation.
【作者单位】：中国测绘科学研究院;武汉大学资源与环境科学学院;
【基金】：基金项目:测绘新技术系统开发与示范应用(2016KJ0104)
【分类号】：X831

【相似文献】