Geoscience Reference
In-Depth Information
Fig. 14.70
Brannerite, − 450 m elevation
Fig. 14.71
Coffinite, − 450 m elevation
14.7.10.2 Grain Size Variables
Histogram and variogram reproduction is heavily influenced
by the secondary variables as there was very little grain size
data. For this reason the histogram and variogram reproduc-
tion for the grain size variables does not exactly match the
input. Moreover, the grain size variables are sparsely sam-
pled suggesting that the input histogram and variogram may
be unreliable. Some deviation from the input parameters due
to the secondary information is warranted.
pendence of the principal components, and (b) influence of
the super secondary attributes on the models.
14.8
Conclusions
Three linear regression models for the prediction of plant
performance from head assay, mineralogy and association
variables. This case study presented a methodology for
the spatial modeling of these variables. The intention is to
use the regression model with the spatial model to predict
plant performance. The cost of obtaining samples of plant
performance (i.e. pilot plant runs) is very high. Building
models based on the sparse sampling of mineral recovery,
14.7.10.3 Association Matrix Variables
There are a total of 100 association variables. Histogram re-
production is not perfect. The resulting histograms and var-
iograms deviate from the input because of (a) lack of inde-
