【摘要】 采用非参数统计理论中的指示克立格与BP神经网络技术,以河套灌区内一典型实验区的土壤水盐预测为例,在一条采样基线上比较了指示克立格与BP神经网络技术的优劣,发现指示克立格与BP神经网络技术在一维空间的土壤水盐预测中能改进普通克立格平滑效应,具有条件模拟的仿真功能,可用于空间水盐插值与预测。更多还原

【Abstract】 Using indicator Kriging in non-parametric approach and BP neural net technique,this paper made the zonal soil water-salt forecasting of a representative experimental section in Hetao irrigation zone as example,contrasted and analyzed two methods in a sampling baseline.It showed that indicator Kriging and BP neural net technique can eliminate glabrous effect of OK method and retrofit conventional GS in threadlike space,they all can be used in spatial water-salt forecasting.更多还原