predict: Spatial model predictions

Description

Make a RasterLayer with a prediction based on a a fitted model object. Provide a Raster* with the independent variables. The layerNames in the RasterStack should exactely match those expected by the model. This will be the case if the RasterStack was used (via extract) to obtain the values to fit the model (see the example). Any type of model (e.g. glm. gam, randomforest) for which a predict method has been implemented can be used.

Usage

predict(object, ...)

Arguments

object

a RasterStack, RasterBrick or RasterLayer object

...

Additional arguments. See under Methods

Value

RasterLayer

Methods

predict(object, model, filename='', fun=predict, ext=NULL, const=NULL, index=1, se.fit=FALSE, na.rm=TRUE, ...) rll{ object a Raster* object. Typcially a multi-layer type (RasterStack or RasterBrick) model A fitted model of any class that has a 'predict' method. E.g. glm, gam, or randomForest filename Output filename for a new raster; if NA the result is not written to a file but returned with the RasterLayer object, in the data slot fun Function. Default value is 'predict', but can be replaced with e.g. predict.se (depending on the type of model) ext An Extent object to limit the prediction to a sub-region of x const data.frame. Can be used to add a constant for which there is no Raster object for model predictions. Particulalry useful if the constant is a character-like factor value for which it is currently not possible to make a RasterLayer index Integer. To select the column if predict.'model' returns a matrix with multiple columns se.fit Compute the standard error AND the predicted value (works for GLM and perhaps other methods where the prediction and se.fit are returned as a list) na.rm Logical. Remove cells with NA values in the predictors before solving the model (and return a NA value for those cells). This option prevents errors with models that cannot handle NA values. In most other cases this will not affect the output. An exception is when predicting with a boosted regression trees model because these return predicted value even if some (or all!) variables are NA ... Additional arguments to pass to the predict.'model' function } The following additional arguments can be passed, to replace default values rll{ format Character. Output file type. See writeRaster datatype Character. Output data type. See dataType overwrite Logical. If TRUE, "filename" will be overwritten if it exists progress Character. "text", "window", or "" (the default, no progress bar) }

Examples

Run this code

# A simple model to predict the location of the R in the R-logo using 20 presence points 
# and 50 (random) pseudo-absence points. This type of model is often used to predict species distributions

# create a RasterStack (a set of predictor rasters)
logo <- stack(system.file("external/rlogo.grd", package="raster"))
layerNames(logo)

#par(mfrow=c(2,2))
#plotRGB(logo, main='logo')
#plot(logo, 1, col=rgb(cbind(0:255,0,0), maxColorValue=255))
#plot(logo, 2, col=rgb(cbind(0,0:255,0), maxColorValue=255))
#plot(logo, 3, col=rgb(cbind(0,0,0:255), maxColorValue=255))

#get presence and random background (pseudo-absence) points
presence <- matrix(c(48, 48, 48, 53, 50, 46, 54, 70, 84, 85, 74, 84, 95, 85, 66, 42, 26, 4, 19, 17, 7, 14, 26, 29, 39, 45, 51, 56, 46, 38, 31, 22, 34, 60, 70, 73, 63, 46, 43, 28), ncol=2)
# random background
background <- cbind(runif(250)*(xmax(logo)-xmin(logo))+xmin(logo), runif(250)*(ymax(logo)-ymin(logo))+ymin(logo))

#extract values for points from stack
xy <- rbind(cbind(1, presence), cbind(0, background))
v <- cbind(xy[,1], extract(logo, xy[,2:3]))
colnames(v)[1] <- 'presback'

#build a model, here an example with glm 
model <- glm(formula=presback~., data=data.frame(v))

#predict to a raster
r <- predict(logo, model, progress='text')

plot(r)
points(presence, bg='blue', pch=21)
points(background, bg='red', pch=21)

## also try:
# require(randomForest)
## formula <- as.factor(presback) ~.
# formula <- presback ~.
# model <- randomForest(formula, data=data.frame(v))
# r2 <- predict(logo, type='response', model, progress='text')
## note the additional argument "type='response'" that is passed to predict.randomForest