RFclass: Random Forest Classification with Random Cross-Validation

In SchoenbergA/IKARUS: Classification Tool

Description

RF Classification with CV

Usage

RFclass(
  tDat,
  predCol = "default",
  predStk = NULL,
  classCol = "class",
  Cores = 1
)

Arguments

tDat	data.frame - with values of the predictors (see details)
predCol	numeric - seq of columns with predictor values. By default uses 1:(length(tDat)-1) for tDat format computed by IKARUS::exrct_Traindat
predStk	RasterStack - with the predictors.
classCol	character - name of the column containing the class information
Cores	numeric - amount of Cores to exclude from calculation, default = 1

Details

• predCol - specific predictors can be selected by setting predCol = x:y

• tDat - the use of IKARUS::exrct_Traindat is recommended.

• parallel processing - the function uses parallel processing for multicore processors. by default all cores -1 are used.

Value

returns a list with the model and the prediction

Author(s)

Andreas Schönberg

Examples

# load data
require(caret)
require(CAST)
require(doParallel)
require(raster)
require(IKARUS)
lau_Stk <- raster::stack(system.file("extdata","lau_RGB.grd",package = "IKARUS"))
lau_tP <-rgdal::readOGR(system.file("extdata","lau_TrainPolygon.shp",package = "IKARUS"))
# handle CRS string
crs(lau_tP) <- crs(lau_Stk)
### extract values using 'exrct_Tdat' to generate training dataset
tDat <- exrct_Traindat(lau_tP,lau_Stk,"class")
# check for class column and predictor columns in input training dataset
head(tDat)
# classification
model1 <- RFclass(tDat = tDat,predCol = "default",predStk = lau_Stk,classCol = "class")
# check model
model1$model_cv

# plot prediction
plot(model1$prediction)

SchoenbergA/IKARUS documentation built on Sept. 8, 2021, 11:11 a.m.