gmGeostats v0.10-6

0

Monthly downloads

0th

Percentile

Geostatistics for Compositional Analysis

Support for geostatistical analysis of multivariate data, in particular data with restrictions, e.g. positive amounts data, compositional data, distributional data, microstructural data, etc. It includes descriptive analysis and modelling for such data, both from a two-point Gaussian perspective and multipoint perspective. The methods mainly follow Tolosana-Delgado, Mueller and van den Boogaart (2018) <doi:10.1007/s11004-018-9769-3>.

Readme

gmGeostats

The goal of gmGeostats is to provide a unified framework for the geostatistical analysis of multivariate data from any statistical scale, e.g. data honoring a ratio scale, or with constraints such as spherical or compositional data.

This R package offers support for geostatistical analysis of multivariate data, in particular data with restrictions, e.g. positive amounts data, compositional data, distributional data, microstructural data, etc. It includes descriptive analysis and modelling for such data, both from a two-point Gaussian perspective and multipoint perspective. The package is devised for supporting 3D, multi-scale and large data sets and grids. This is a building block of the suite of HIF geometallurgical software.

Installation

You can install the released version of gmGeostats from CRAN with:

install.packages("gmGeostats")

Example

Read the package vignette for an extended scheme of the package functionality. The fundamental steps are:

## load the package and its dependencies
library(gmGeostats)
#> Welcome to 'gmGeostats', a package for multivariate geostatistical analysis.
#>  Note: use 'fit_lmc' instead of fit.lmc

## read your data, identify coordinates and sets of variables
data("Windarling") # use here some read*(...) function
colnames(Windarling)
#>  [1] "Hole_id"     "Sample.West" "Sample.East" "West"        "East"       
#>  [6] "Easting"     "Northing"    "Lithotype"   "Fe"          "P"          
#> [11] "SiO2"        "Al2O3"       "S"           "Mn"          "CL"         
#> [16] "LOI"
X = Windarling[,c("Easting", "Northing")]
Z = Windarling[,c(9:12,14,16)]

## declare the scale of each set of variables
Zc = compositions::acomp(Z) # other scales will come in the future

## pack the data in a gmSpatialModel object using an appropriate
#     make.** function
gsm = make.gmCompositionalGaussianSpatialModel(
  data = Zc, coords = X, V = "alr", formula = ~1
)

From this point on, what you do depends on which model do you have in mind. Here we briefly cover the case of a Gaussian model, though a multipoint approach can also be tackled with function make.gmCompositionalMPSSpatialModel() providing a training image as model. See the package vignette for details.

A structural analysis can be obtained in the following steps

## empirical structural function
vge = variogram(gsm)

## model specification
vm = gstat::vgm(model="Sph", range=25, nugget=1, psill=1)
# you can use gstat specifications!

## model fitting
gsm.f = fit_lmc(v = vge, g = gsm, model = vm)

## plot
variogramModelPlot(vge, model = gsm.f)

This model can then be validated, interpolated and/or simulated. The workflow for each of these tasks is always:

1.- define some method parameters with a tailored function, e.g. LeaveOneOut() for validation, KrigingNeighbourhood() for cokriging or SequentialSimulation() for sequential Gaussian Simulation

2.- if desired, define some new locations where to interpolate or simulate, using expand.grid() or sp::GridTopology() or similar

3.- call an appropriate function, specifying the model, potential new data, and the parameters created in the preceding steps; e.g. validate(model, pars) for validation, or predict(model, newdata, pars) for interpolation or simulation

More information can be found in the package vignette.

Functions in gmGeostats

Name Description
ModelStructuralFunctionSpecification-class Structural function model specification
LeaveOneOut Specify the leave-one-out strategy for validation of a spatial model
GridOrNothing-class Superclass for grid or nothing
CholeskyDecomposition Create a parameter set specifying a LU decomposition simulation algorithm
Maf Generalised diagonalisations Calculate several generalized diagonalisations out of a data set and its empirical variogram
DataFrameStack Create a data frame stack
LMCAnisCompo Create a anisotropic model for regionalized compositions
DSpars Create a parameter set specifying a direct sampling algorithm
EmpiricalStructuralFunctionSpecification-class Empirical structural function specification
KrigingNeighbourhood Create a parameter set of local for neighbourhood specification.
Windarling Ore composition of a bench at a mine in Windarling, West Australia.
SequentialSimulation Create a parameter set specifying a gaussian sequential simulation algorithm
ana Flow anamorphosis transform Compute a transformation that gaussianizes a certain data set
as.array.DataFrameStack Convert a stacked data frame into an array
anaBackward Backward gaussian anamorphosis backward transformation to multivariate gaussian scores
as.LMCAnisCompo Recast compositional variogram model to format LMCAnisCompo
as.gstatVariogram Represent an empirical variogram in "gstatVariogram" format
NfoldCrossValidation Specify a strategy for validation of a spatial model
as.gstat Convert a regionalized data container to gstat
NGSAustralia National Geochemical Survey of Australia: soil data
anaForward Forward gaussian anamorphosis forward transformation to multivariate gaussian scores
as.list.DataFrameStack Convert a stacked data frame into a list of data.frames
accuracy Compute accuracy and precision
anis2D.par2A Produce anisotropy matrix from angle and anisotropy ratios
TurningBands Create a parameter set specifying a turning bands simulation algorithm
as.AnisotropyScaling Convert to anisotropy scaling matrix
as.logratioVariogram Recast empirical variogram to format logratioVariogram
as.gmEVario Convert empirical structural function to gmEVario format
as.gmSpatialModel Recast spatial object to gmSpatialModel format
as.function.gmCgram Convert a gmCgram object to an (evaluable) function
getStackElement Set or get the i-th data frame of a data.frame stack
as.gmCgram.variogramModelList Convert theoretical structural functions to gmCgram format
as.variogramModel Convert an LMC variogram model to gstat format
coloredBiplot.genDiag Colored biplot for gemeralised diagonalisations Colored biplot method for objects of class genDiag
gmGaussianSimulationAlgorithm-class parameters for Gaussian Simulation methods
gmSpatialModel-class Conditional spatial model data container
gmTrainingImage-class MPS training image class
plot.gmCgram Draw cuves for covariance/variogram models
plot.accuracy Plot method for accuracy curves
is.isotropic Check for anisotropy of a theoretical variogram
gsi.gstatCokriging2compo Reorganisation of cokriged compositions
precision Precision calculations
gsi.orig extract information about the original data, if available
length.gmCgram Length, and number of columns or rows
predict.LMCAnisCompo Compute model variogram values Evaluate the variogram model provided at some lag vectors
constructMask Constructs a mask for a grid
as.CompLinModCoReg Recast a model to the variogram model of package "compositions"
getMask Get the mask info out of a spatial data object
gmValidationStrategy-class Validation strategy description
gmSimulationAlgorithm-class Parameter specification for a spatial simulation algorithm
has.missings.data.frame Check presence of missings check presence of missings in a data.frame
gsi.produceV Create a matrix of logcontrasts and name prefix
gmUnconditionalSpatialModel-class General description of a spatial model
gmNeighbourhoodSpecification-class Neighbourhood description
gmMPSParameters-class parameters for Multiple-Point Statistics methods
fit_lmc.logratioVariogramAnisotropy Fit an LMC to an empirical variogram
dimnames.DataFrameStack Return the dimnames of a DataFrameStack
logratioVariogram,acomp-method Logratio variogram of a compositional data
gsi.CondTurningBands Internal function, conditional turning bands realisations
gsi.Cokriging Cokriging of all sorts, internal function
noSpatCorr.test Test for lack of spatial correlation
pairsmap Multiple maps Matrix of maps showing different combinations of components of a composition, user defined
logratioVariogram Empirical logratio variogram calculation
gmApply Apply Functions Over Array or DataFrameStack Margins
getTellus Download the Tellus survey data set (NI)
plot.gmEVario Plot empirical variograms
gmSpatialDataContainer-class General description of a spatial data container
gmSpatialMethodParameters-class Parameter specification for any spatial method
gsi.DS Workhorse function for direct sampling
make.gmMultivariateGaussianSpatialModel Construct a Gaussian gmSpatialModel for regionalized multivariate data
plot.logratioVariogramAnisotropy Plot variogram lines of empirical directional logratio variograms
spatialGridRmult Construct a regionalized multivariate data
spectralcolors Spectral colors palette based on the RColorBrewer::brewer.pal(11,"Spectral")
gmGaussianMethodParameters-class parameters for Spatial Gaussian methods of any kind
setCgram gmGeostats Variogram models set up a D-variate variogram models
gsi.TurningBands Internal function, unconditional turning bands realisations
setGridOrder Set or get the ordering of a grid
gsi.calcCgram Compute covariance matrix oout of locations
stackDim Get/set name/index of (non)stacking dimensions
mean.accuracy Mean accuracy
[.DataFrameStack Extract rows of a DataFrameStack
xvErrorMeasures Cross-validation errror measures
image.logratioVariogramAnisotropy Plot variogram maps for anisotropic logratio variograms
[.gmCgram Subsetting of gmCgram variogram structures
predict.gmSpatialModel Predict method for objects of class 'gmSpatialModel'
stackDim,Spatial-method Get name/index of the stacking dimension of a Spatial object
predict.genDiag Predict method for generalised diagonalisation objects
[[.gmCgram Subsetting of gmCgram variogram structures
swarmPlot Plot a swarm of calculated output through a DataFrameStack
[.logratioVariogramAnisotropy Subsetting of logratioVariogram objects
sphTrans Spherifying transform Compute a transformation that spherifies a certain data set
gsi.EVario2D Empirical variogram or covariance function in 2D
image_cokriged Plot an image of gridded data
variogramModelPlot.gstatVariogram Quick plotting of empirical and theoretical variograms Quick and dirty plotting of empirical variograms/covariances with or without their models
image.mask Image method for mask objects
is.anisotropySpecification Check for any anisotropy class
variogramModelPlot.logratioVariogram Quick plotting of empirical and theoretical logratio variograms Quick and dirty plotting of empirical logratio variograms with or without their models
ndirections Number of directions of an empirical variogram
mean.spatialDecorrelationMeasure Average measures of spatial decorrelation
setMask Set a mask on an object
print.mask Print method for mask objects
pwlrmap Compositional maps, pairwise logratios Matrix of maps showing different combinations of components of a composition, in pairwise logratios
sortDataInGrid Reorder data in a grid
validate Validate a spatial model
variogramModelPlot Quick plotting of empirical and theoretical variograms Quick and dirty plotting of empirical variograms/covariances with or without their models
make.gmCompositionalGaussianSpatialModel Construct a Gaussian gmSpatialModel for regionalized compositions
make.gmCompositionalMPSSpatialModel Construct a Multi-Point gmSpatialModel for regionalized compositions
plot.swarmPlot Plotting method for swarmPlot objects
+.gmCgram Combination of gmCgram variogram structures
spatialGridAcomp Construct a regionalized composition / reorder compositional simulations
swath Swath plots
spatialDecorrelation Compute diagonalisation measures
logratioVariogram_gmSpatialModel Variogram method for gmSpatialModel objects
write.GSLib Write a regionalized data set in GSLIB format
unmask Unmask a masked object
No Results!

Vignettes of gmGeostats

Name
gmGeostats.Rmd
No Results!

Last month downloads

Details

Date 2020-09-07
License CC BY-SA 4.0 | GPL (>= 2)
URL https://www.r-project.org, http://www.stat.boogaart.de
Copyright (C) 2020 by Helmholtz-Zentrum Dresden-Rossendorf and Edith Cowan University; gsi.DS code by Hassan Talebi
RoxygenNote 7.1.0
Encoding UTF-8
VignetteBuilder knitr
LazyData true
Collate 'Anamorphosis.R' 'compositionsCompatibility.R' 'gstatCompatibility.R' 'variograms.R' 'gmValidationStrategy.R' 'gmAnisotropy.R' 'abstractClasses.R' 'accuracy.R' 'data.R' 'exploratools.R' 'genDiag.R' 'geostats.R' 'gmDataFrameStack.R' 'gmSimulation.R' 'gmSpatialMethodParameters.R' 'gmSpatialModel.R' 'grids.R' 'mask.R' 'spSupport.R' 'uncorrelationTest.R' 'zzz.R'
NeedsCompilation yes
Packaged 2020-09-08 13:43:49 UTC; tolosa53
Repository CRAN
Date/Publication 2020-09-16 09:10:03 UTC

Include our badge in your README

[![Rdoc](http://www.rdocumentation.org/badges/version/gmGeostats)](http://www.rdocumentation.org/packages/gmGeostats)