If you have access to OneMine as part of a member benefit, log in through your member association website for a seamless user experience.
|Qualitative and interpretive information as obtained from geological and/or geophysical studies, in addition to precise measurement, is incorporated into the estimation process through an indicator variable. The indicator conditioned estimator is formulated as a cokriging, and its expression is data value dependent as opposed to that of the traditional kriging estimators. The solution of the dual cokriging system provides simultaneously an estimate of the unknown and an estimate of the conditional probability that the unknown belongs to a specific population or class. Being an unbiased and minimum-error-variance estimator by construction, the indicator conditioned estimator is better in the least square sense than the ordinary kriging estimator because it uses additional information carried by the indicator data. A case study using a large reference database is presented; the indicator conditioned estimates have a smaller conditional bias and a smaller error variance than the ordinary kriging estimates. The results also showed that the indicator conditioned estimator is more efficient for selection in mining.|