biomod2: sample.factor.levels – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

sample.factor.levels

Tool to ensure the sampling of all levels of a factorial variable

Description

This function will sample randomly an element of each level of all the factorial variables contains in a Raster* object or a data.frame

Usage

sample.factor.levels(x, mask.out = NULL, mask.in = NULL)

Arguments

`x`	a Raster* object or a data.frame
`mask.out`	a Raster/data.frame mask containing area that have already been sampled. The factor levels within this mask will not be sampled.
`mask.in`	a Raster/list of Raster/data.frame mask (potentially a stack of masks) indicating areas were we want to sample our factor level in priority. Note that if after having explore this masks some levels of the considered factorial varialble remains unsampled, this levels will be sampled in the reference input object (here 'x')

Details

In case any factorial variable is found in the input object then NULL is returned.

Value

a numeric vector the number (cell number for Raster* objects or row number for data.frame) where each will refer to a single level of a single factorial variable.

Note

- The x/mask.out/mask.in should be coherent in term of dimention (same number of rows for data.frame and same number of rows, column, identic resolution and projection coordinates system for Raster* objects) - If mask.in contains several masks (RasterStack or multi-column data.frame) then the order of the mask matter. The mask will be considered successively. The first will be use prioritarly to sample our variable factor levels and so on. - Raster* masks will be understood as: - NA: out of of mask - not NA: in mask - data.frame masks will be understood as: - FALSE: out of mask - TRUE: in mask

Author(s)

damien g.

Examples

## example with raster* object ---------- 
library(raster)
## create a factorial raster
r1 <- raster()
r1[] <- 1; r1[1] <- 2; r1[2:3] <- 3
r1 <- as.factor(r1)
## create a continuous raster
r2 <- raster()
r2[] <- rnorm(ncell(r2))
## pull the raster into a RasterStack
stk <- stack(r1, r2)
is.factor(stk)

## define a mask for already sampled points
mask.out <- r1
mask.out[] <- NA; mask.out[2:3] <- 1

## define a list of mask where we want to sample in priority
mask.in.1 <- mask.in.2 <- r1
mask.in.1[1:10] <- NA ## only level 1 should be sampled in this mask
mask.in.2[1] <- NA ## only levels 1 and 3 should be sampled in this mask
mask.in <- list(mask.in.1 = mask.in.1, 
                mask.in.2 = mask.in.2)

## test different version of the function
sample.factor.levels(stk, mask.out = mask.out)
sample.factor.levels(stk, mask.in = mask.in)
sample.factor.levels(stk, mask.out = mask.out, mask.in = mask.in)

biomod2

Ensemble Platform for Species Distribution Modeling

v3.4.6

GPL-2

Authors

Wilfried Thuiller [aut, cre], Damien Georges [aut, cre], Robin Engler [aut], Frank Breiner [aut]

Initial release

2020-02-24