lmc: Linear Model of Corregionalization

In victorvicpal/MGSR: Multivariate Gaussian Subspatial Regression

Description

Linear Model of Corregionalization fitting

Usage

lmc(CV, fun, a, tol = 0.001, mode = "aut")

Arguments

CV	Crossvariogram resulting from crossvariogram
fun	Vector of functions. Use 'Sph' for spherical, 'Gau' for Gaussian, 'Exp' for Exponential, 'Pow' for Power and 'Lin' for lineal.
a	Vector of ranges. Same length as 'fun'. If you use the Power function which is more suitable for subspatial fitting (0<a<2).
tol	Tolerance
mode	Automatic 'aut' or Manual 'man'. Manual mode allows for chosing sill values individually.

Details

Power distribution is suitable only for small distances.

Value

Object containing sill matrices and WSS results.

Note

Combination of multiple ranges and functions are also allowed.

Author(s)

Victor Vicente Palacios

References

Pelletier B, Dutilleul P, Larocque G, Fyles J.W. Fitting the Linear Model of Coregionaliza- tion by Generalized Least Squares. Math Geol. 2004;36(3):323-343.

See Also

crossvariogram,plot.crossvariogram

Examples

data("iris")
Versicolor <- iris[which(iris$Species=='versicolor'),-5]
##Data Standarization
means_vers <- apply(Versicolor,2,mean)
sd_vers <- apply(Versicolor,2,sd)
Versicolor_st <- Versicolor

for (i in 1:length(Versicolor[1,]))
{Versicolor_st[,i] <- (Versicolor[,i]-means_vers[i])/sd_vers[i]}

##PrComp
PC_train <- princomp(Versicolor_st)
## CrossVariogram Calculation
CV_vers <- crossvariogram(as.data.frame(PC_train$scores[,1:2]),as.data.frame(Versicolor_st),11)
## lmc 
RES_vers <- lmc(CV_vers,'Pow',1.6)
## plot
plot.crossvariogram(CV_vers,RES_vers)

victorvicpal/MGSR documentation built on May 3, 2019, 6:11 p.m.