变异函数模型参数估计的信息熵加权回归法

PARAMETER ESTIMATION OF VARIOGRAM MODEL BY USING INFORMATION ENTROPY WEIGHTED REGRESSION

将熵权理论引入到变异函数理论模型的参数估计中,对加权多项式回归的加权方法进行改进,以信息熵定权,并用实际变形监测数据进行实验。实验结果表明,该方法综合了距离和样本点对数对权重的影响,加权回归确定的变异函数模型更准确,插值预报效果更优。

The information entropy theory was introduced into parameter estimation of variogram model, to improve weighting method of weighted polynomial regression using the entropy weight method. An experiment was taken with real deformation monitoring data, considering the influence of both distance and the number of point pairs at the same distance. The experimental results show that the variogram model determined by information entropy weighted regression is more reasonable, and interpolating prediction is more accurate.