模糊多重指示克里格方法的探讨与运用

Discussion on Fuzzy Multiple Indicator Kriging with a Case Study

为探寻矿产资源量估值更精确的技术方法,运用模糊多重指示克里格方法,探求降低高偏态分布数据对于矿产资源估值的影响。目前,虽然有一系列技术方法用于建模和表征高偏态分布数据,如指示克里格法。但是这些传统的指示方法需要以规律的区间在整个样本值分布上,设定一组阀值下的指示转换。在某个阀值区间中,所有不同品位的数据均被赋予相同的二进制转换代码(0,1),其估值结果多显示局部的精确度,仍然有待于改进。介绍了模糊理论原理,提出了模糊多重指示克里格方法,其中的标准二进制指示代码由一个模糊转换隶属函数所替代,因而样本值与某一阈值的接近程度被模糊转换隶属函数所量化。研究表明:模糊多重指标克里格方法能够有效降低全局条件偏差,增进局部精确性,具有进一步研讨的意义。

When estimating a mineral resource,it is important to obtain accurate local grade estimates and local estimation methods such as ordinary and indicator kriging are now in widespread use.Both distribution skewness and grade outliers influence the distribution in this area,and a more sophisticated method of mean grade selection is required if conditional bias is to be avoided,or at least substantially reduced.In this paper,the objective is to investigate a fuzzy indicator kriging approach in which the standard binary indicator coding is replaced by a coding via a symmetric fuzzy transfer membership function,which incorporates a measure of the proximity of the sample data to the chosen thresholds.The author will compare the method with both ordinary and standard multiple indicator kriging.The method of fuzzy indicator kriging presented here shows that it reduces the global conditional bias and local accuracy relative to both ordinary point kriging and traditional indicator kriging.