crossvalidate: This function performs crossvalidation in the style of...

Description Usage Arguments Value Note Author(s) Examples

View source: R/crossvalidate.r

Description

In addition to supporting kriging and idw, it also supports thin plate splines (through Tps (fields)).

Usage

1
crossvalidate(formula, data, func = "doTps", ...)

Arguments

formula

A formula to define the dependent and the independent variables, see the documentation of krige.

data

The input data, should be a spatial object which supports coordinates extracting through coordinates()

func

The function which should be used for cross-validation, possibilities are "krige", "idw", "doNearestNeighbor" and "doTps".

...

parameters that are passed on to 'func'

Value

The function returns a spatial object with the class of 'data'. The format in which the cross-validation results are presented is equal to that of krige.cv.

Note

At this stage only leave-one-out cross-validation is supported. In addition, the formula supports only a limit syntax, i.e. log(zinc) or sqrt(dist) are not possible in the formula. Instead, create a new column in the object, e.g. meuse$log_zinc = log(meuse$zinc).

Author(s)

Paul Hiemstra, p.h.hiemstra@gmail.com

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
library(automap)
library(sp)
library(gstat)
data(meuse)
coordinates(meuse) = ~x+y
data(meuse.grid)
gridded(meuse.grid) = ~x+y

regres_cv = crossvalidate(zinc~dist, meuse, func = "krige", debug.level = 0)
nearestneighbor_cv = crossvalidate(zinc~1, meuse, func = "doNearestNeighbor")
idw2_cv = crossvalidate(zinc~1, meuse, func = "idw", debug.level = 0)
idw4_cv = crossvalidate(zinc~1, meuse, func = "idw", debug.level = 0, idp = 4)
idw05_cv = crossvalidate(zinc~1, meuse, func = "idw", debug.level = 0, idp = 0.5)
ked_cv = crossvalidate(zinc~dist, meuse, func = "krige", 
        model = autofitVariogram(zinc~dist, meuse, model = "Ste")$var_model, debug.level = 0)
tps_cv = crossvalidate(zinc~1, meuse, func = "doTps", debug.level = 0)
tpsdist_cv = crossvalidate(zinc~dist, meuse, func = "doTps", debug.level = 0)

compare.cv(regres_cv, idw2_cv, idw4_cv, idw05_cv, tps_cv, tpsdist_cv, ked_cv)

MariekeDirk/GeoInterpolation documentation built on May 14, 2019, 8:20 a.m.