estimateParameters.psgp {psgp}

Parameter estimation using a Projected Sequential Gaussian Process (PSGP)

Description

This overloads the estimateParameters routine from the intamap package for interpolation using the PSGP method.

Usage

  estimateParameters(object, ...)

Arguments

object

a list object of Intamap type. Most arguments necessary for interpolation are passed through this object. See intamap-package for further description of the necessary content of this variable.

...

other parameters for the generic method, not used for this method

Details

See psgp-package and learnParameters for further details.

Author(s)

Remi Barillec, Ben Ingram

See Also

learnParameters, estimateParameters, makePrediction, createIntamapObject

Examples

# load our favourite dataset
data(meuse)
coordinates(meuse) = ~x+y
meuse$value = log(meuse$zinc)
data(meuse.grid)
gridded(meuse.grid) = ~x+y
proj4string(meuse) = CRS("epsg:28992")
proj4string(meuse.grid) = CRS("epsg:28992")

# the following two steps are only needed if one wishes to
# include observation errors

# indicate which likelihood model should be used for each observation
# in this case we use a different model for each observation
nobs = length(meuse$value)          # Number of observations
meuse$oeid  <- seq(1:nobs)
  
# the variances for the error models are random in this example
# in real examples they will come from actual measurements 
# characteristics
meuse$oevar <- abs( rnorm( max(meuse$oeid) ) )

# set up intamap object:
obj = createIntamapObject(
  observations = meuse,
  predictionLocations = meuse.grid,
  targetCRS = "epsg:3035",
  class = "psgp"    # Use PSGP for parameter estimation/interpolation
)

# do interpolation step:
obj = conformProjections(obj)
obj = estimateParameters(obj)