GeoModels (version 2.2.6)

GeoPit: Probability integral or normal score transformation

Description

For a given GeoFit, the procedure applies either the probability integral transformation or the normal score transformation to the data.

Usage

GeoPit(object, type = "Uniform")

Value

Returns an (updated) object of class GeoFit

Arguments

object

An object of class GeoFit, GeoKrig, or GeoKrigloc. For fitted models with covariates, the GeoFit object must first be transformed by GeoResiduals.

type

The type of transformation. If "Uniform", the probability integral transformation is performed. If "Gaussian", the normal score transformation is performed.

Author

Moreno Bevilacqua, ,https://sites.google.com/view/moreno-bevilacqua/home, Víctor Morales Oñate, , https://sites.google.com/site/moralesonatevictor/, Christian Caamaño-Carrillo, ,https://www.researchgate.net/profile/Christian-Caamano

Examples

library(GeoModels)

model="Beta2"
copula="Clayton"

set.seed(221)
NN=800
x <- runif(NN);y <- runif(NN)
coords=cbind(x,y)


shape=1.5
scale=0.2;power2=4
smooth=0
nugget=0
nu=8

corrmodel="GenWend"

min=-2;max=1
mean=0


param=list(smooth=smooth,power2=power2, min=min,max=max,
 mean=mean, nu=nu,
 scale=scale,nugget=nugget,shape=shape)

optimizer="nlminb"

data <- GeoSimCopula(coordx=coords, corrmodel=corrmodel, 
model=model,param=param,copula=copula)$data

I=50
fixed<-list(nugget=nugget,sill=1,scale=scale,smooth=smooth,power2=power2,min=min,max=max,nu=nu)
start<-list(shape=shape,mean=mean)
lower<-list(shape=0,mean=-I)
upper<-list(shape=10,mean=I)

#### maximum independence likelihood
fit1 <- GeoFit(data=data,coordx=coords,corrmodel=corrmodel,
model=model,likelihood="Marginal",type="Independence",
 optimizer=optimizer,lower=lower,
 upper=upper,copula=copula,
 start=start,fixed=fixed)

## PIT transformation
aa=GeoPit(fit1,type="Uniform")
hist(aa$data,freq=FALSE)
GeoScatterplot(aa$data,coords,neighb=c(1,2))
## Normal score transformation
bb=GeoPit(fit1,type="Gaussian")
hist(bb$data,freq=FALSE)
GeoScatterplot(bb$data,coords,neighb=c(1,2))