Kriging is an estimation method in which the accuracy of the estimate depends on several factors,
namely:
1. Number of samples and quality of data at each point
2. Sample position. Equally spaced samples achieve good coverage and provide more
information about the deposit
3. The distance between samples and points or blocks to be estimated. Prioritize adjacent
samples
4. The spatial continuity of the variables under consideration. It is easier to estimate the
value of a sufficiently regular variable than one that fluctuates randomly.
If the N data values are known from z(x1),…z(xn) and the linear function of the variable Z(x) will be
estimated. For example at point Z(x0) or the average in a certain area. Several other linear functions
such as the gradient can also be said to be kriging estimates. Then it can be defined:
Volume V can be an entire deposit or mining block or a point in a point estimate or even an irregular
shape.
The regular shape makes it easier to choose the point size or the number of points in the zone to
estimate. This is more difficult for irregular shapes. As shown in the following image, a slight change in
grid spacing will change the nodes within the zone
To estimate Z(v) , the average data weight is:
where λi is the weight factor. By convention, ' star ' will be used to denote an estimated value as
opposed to an actual but unknown value. The problem is to choose the weighting factor in the best way,
namely the distance or spacing of the geostatistical model samples. With regional variables ie
namely:
1. Number of samples and quality of data at each point
2. Sample position. Equally spaced samples achieve good coverage and provide more
information about the deposit
3. The distance between samples and points or blocks to be estimated. Prioritize adjacent
samples
4. The spatial continuity of the variables under consideration. It is easier to estimate the
value of a sufficiently regular variable than one that fluctuates randomly.
If the N data values are known from z(x1),…z(xn) and the linear function of the variable Z(x) will be
estimated. For example at point Z(x0) or the average in a certain area. Several other linear functions
such as the gradient can also be said to be kriging estimates. Then it can be defined:
Volume V can be an entire deposit or mining block or a point in a point estimate or even an irregular
shape.
The regular shape makes it easier to choose the point size or the number of points in the zone to
estimate. This is more difficult for irregular shapes. As shown in the following image, a slight change in
grid spacing will change the nodes within the zone
To estimate Z(v) , the average data weight is:
where λi is the weight factor. By convention, ' star ' will be used to denote an estimated value as
opposed to an actual but unknown value. The problem is to choose the weighting factor in the best way,
namely the distance or spacing of the geostatistical model samples. With regional variables ie
