【摘要】 克里金插值法根据待测点与已知点的空间位置关系以及研究区域内变量的空间相关性为已知点赋予不同权重，再通过加权求和的方式得到待测点的估计值，为待测点提供最佳的线性无偏估计。变异函数描述了区域化变量的空间变化信息，常用的理论变异函数包括球状模型、指数模型和高斯模型，然而以上模型形式较为固定且选择较为主观，可能无法准确反映数据波动趋势。基于此，提出引入支持向量机作为新的变异函数对克里金方法进行改进。使用汾渭平原PM_2.5浓度数据对改进前后的克里金插值法进行效果对比。结果表明：支持向量机能够拟合出变异函数的变化趋势，改进后的克里金插值精度优于改进前的克里金插值精度，因此支持向量机改进的克里金插值法是一种可选的克里金方法。更多还原

【Abstract】 The Kriging interpolation method assigns different weights to the known points according to the spatial position relationship between the points to be measured and the spatial correlation of the variables in the study area, and then obtains the estimated value of the points to be measured by weighted summation, so as to provide the best linear unbiased estimation of the points to be measured. Variogram describes the spatial variation information of regionalized variables, and commonly used theoretical variograms include spherical models, exponential models, and Gaussian models, but the above model forms are more fixed and the selection is more subjective, which may not accurately reflect the data fluctuation trend. Based on this, it was proposed to introduce a support vector machine as a new variogram to improve the kriging method. The PM_2.5 concentration data of the Fenwei Plain was used to compared the effects of Kriging interpolation before and after improvement. The results show that the support vector machine can fit the variation trend of the variogram, the improved Kriging interpolation accuracy is better than the Kriging interpolation accuracy before the improvement, so the improved Kriging interpolation method of the support vector machine is an optional Kriging method.更多还原