Description Usage Arguments Details References See Also
Implicit geological using Gaussian Processes and compositional relations among the geological classes.
1 2 3 | SPGP_geomod(data, value1, value2 = value1, model, nugget, tangents = NULL,
pseudo_inputs, pseudo_tangents = NULL, variational = T,
enforce.contacts = F, reg.v = 1e-09, reg.t = 1e-09)
|
data |
A 3D spatial object. |
value1 |
A column name indicating the variable that contains the geological classes. |
value2 |
Another column name, with possibly different class labels. |
model |
The covariance model. A |
nugget |
The model's nugget effect or noise variance. |
tangents |
A |
pseudo_inputs |
The desired number of pseudo-inputs (whose coordinates will be sampled from the data) or a matrix or data frame with their coordinates. |
pseudo_tangents |
The desired number of pseudo-structural data (whose
coordinates will be sampled from the data) or a |
variational |
Use the variational approach? |
enforce.contacts |
Force the model to interpolate geological contacts? |
reg.v |
Regularization to improve stability. A single value or a vector with length matching the number of data points. |
reg.t |
Regularization for structural data. A single value or a vector with length matching the number of structural data. |
The role of this class is to manage multiple SPGP
objects, one
per geological class in the data, that model an indicator variable or
"potential" for each class. A point in 3D space is assigned to the class
with the highest modeled potential. The class labels are defined as
unique(c(data[[value1]], data[[value2]]))
.
The points at which data[[value1]] != data[[value2]]
are considered
to lie exactly at a boundary between two geological classes.
If enforce.contacts = T
the force.interp
option of the GP
class is turned on in order
to make the geological boundaries pass through them.
The prediction provides an indicator value for each class at each location,
as well as the label of the most likely class. In locations away from the
data, the most likely class can be an extra, "Unknown"
class. This
class is there to to ensure that the probabilities sum to 1 according to
compositional data principles, and also as a measure of model uncertainty.
The geological boundaries can be drawn by contouring each class's indicator at the value 0.
Gon<c3><a7>alves <c3><8d>G, Kumaira S, Guadagnin F. A machine learning approach to the potential-field method for implicit modeling of geological structures. Comput Geosci 2017;103:173<e2><80><93>82. doi:10.1016/j.cageo.2017.03.015.
SPGP
, SPGP-init
, Predict
,
Make3DArray
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.