randomForestSRC (version 2.5.0)

vimp: VIMP for Single or Grouped Variables

Description

Calculate variable importance (VIMP) for a single variable or group of variables for training or test data.

Usage

# S3 method for rfsrc
vimp(object, xvar.names, outcome.target=NULL, 
  importance = c("permute", "random", "anti",
                  "permute.ensemble", "random.ensemble", "anti.ensemble"),
  joint = FALSE, subset, seed = NULL, do.trace = FALSE, ...)

Arguments

object

An object of class (rfsrc, grow) or (rfsrc, forest). Requires forest=TRUE in the original rfsrc call.

xvar.names

Names of the x-variables to be used. If not specified all variables are used.

outcome.target

Character vector for multivariate families specifying the target outcomes to be used. The default is to use all coordinates.

importance

Type of VIMP.

joint

Individual or joint VIMP?

subset

Vector indicating which rows of the grow data to restrict VIMP calculations to; i.e. this option yields VIMP which is restricted to a specific subset of the data. Note that the vector should correspond to the rows of object$xvar and not the original data passed in the grow call. All rows used if not specified.

seed

Negative integer specifying seed for the random number generator.

do.trace

Number of seconds between updates to the user on approximate time to completion.

...

Further arguments passed to or from other methods.

Value

An object of class (rfsrc, predict), which is a list with the following key components:

err.rate

OOB error rate for the ensemble restricted to the subsetted data.

importance

Variable importance (VIMP).

Details

Using a previously grown forest, calculate the VIMP for variables xvar.names. By default, VIMP is calculated for the original data, but the user can specify a new test data for the VIMP calculation using newdata. Depending upon the option importance, VIMP is calculated either by random daughter assignment or by random permutation of the variable(s). The default is Breiman-Cutler permutation VIMP. See rfsrc for more details.

Joint VIMP is requested using joint. The joint VIMP is the importance for a group of variables when the group is perturbed simultaneously.

References

Ishwaran H. (2007). Variable importance in binary regression trees and forests, Electronic J. Statist., 1:519-537.

See Also

rfsrc

Examples

# NOT RUN {
## ------------------------------------------------------------
## classification example
## showcase different vimp
## ------------------------------------------------------------

iris.obj <- rfsrc(Species ~ ., data = iris)

# Breiman-Cutler permutation vimp
print(vimp(iris.obj)$importance)

# Breiman-Cutler random daughter vimp
print(vimp(iris.obj, importance = "random")$importance)

# Breiman-Cutler joint permutation vimp 
print(vimp(iris.obj, joint = TRUE)$importance)

# Breiman-Cuter paired vimp
print(vimp(iris.obj, c("Petal.Length", "Petal.Width"), joint = TRUE)$importance)
print(vimp(iris.obj, c("Sepal.Length", "Petal.Width"), joint = TRUE)$importance)


## ------------------------------------------------------------
## regression example
## compare Breiman-Cutler vimp to ensemble based vimp
## ------------------------------------------------------------

airq.obj <- rfsrc(Ozone ~ ., airquality)
vimp.all <- cbind(
     ensemble = vimp(airq.obj, importance = "permute.ensemble")$importance,
     breimanCutler = vimp(airq.obj, importance = "permute")$importance)
print(vimp.all)


## ------------------------------------------------------------
## regression example
## calculate VIMP on test data
## ------------------------------------------------------------

set.seed(100080)
train <- sample(1:nrow(airquality), size = 80)
airq.obj <- rfsrc(Ozone~., airquality[train, ])

#training data vimp
print(airq.obj$importance)
print(vimp(airq.obj)$importance)

#test data vimp
print(vimp(airq.obj, newdata = airquality[-train, ])$importance)

## ------------------------------------------------------------
## survival example
## study how vimp depends on tree imputation
## makes use of the subset option
## ------------------------------------------------------------

data(pbc, package = "randomForestSRC")

# determine which records have missing values
which.na <- apply(pbc, 1, function(x){any(is.na(x))})

# impute the data using na.action = "na.impute"
pbc.obj <- rfsrc(Surv(days,status) ~ ., pbc, nsplit = 3,
        na.action = "na.impute", nimpute = 1)

# compare vimp based on records with no missing values
# to those that have missing values
# note the option na.action="na.impute" in the vimp() call
vimp.not.na <- vimp(pbc.obj, subset = !which.na, na.action = "na.impute")$importance
vimp.na <- vimp(pbc.obj, subset = which.na, na.action = "na.impute")$importance
print(data.frame(vimp.not.na, vimp.na))
# }