caret (version 4.47)

rfe: Backwards Feature Selection

Description

A simple backwards selection, a.k.a. recursive feature selection (RFE), algorithm

Usage

rfe(x, ...)
## S3 method for class 'default':
rfe(x, y, 
    sizes = 2^(2:4), 
    metric = ifelse(is.factor(y), "Accuracy", "RMSE"),
    maximize = ifelse(metric == "RMSE", FALSE, TRUE),
    rfeControl = rfeControl(), 
    ...)
rfeIter(x, y, 
        testX, testY, 
        sizes, 
        rfeControl = rfeControl(), 
        ...)

Arguments

x

a matrix or data frame of predictors for model training. This object must have unique column names.

y

a vector of training set outcomes (either numeric or factor)

testX

a matrix or data frame of test set predictors. This must have the same column names as x

testY

a vector of test set outcomes

sizes

a numeric vector of integers corresponding to the number of features that should be retained

metric

a string that specifies what summary metric will be used to select the optimal model. By default, possible values are "RMSE" and "Rsquared" for regression and "Accuracy" and "Kappa" for classification. If custom performance metrics are used (via the

maximize

a logical: should the metric be maximized or minimized?

rfeControl

a list of options, including functions for fitting and prediction. See the package vignette or rfeControl for examples

...

options to pass to the model fitting function

Value

• A list with elements
• finalVariables
  a list of size length(sizes) + 1 containing the column names of the ``surviving'' predictors at each stage of selection. The first element corresponds to all the predictors (i.e. size = ncol(x))
• pred
  a data frame with columns for the test set outcome, the predicted outcome and the subset size.

Details

This function implements backwards selection of predictors based on predictor importance ranking. The predictors are ranked and the less important ones are sequentially eliminated. The package vignette for feature selection has detailed descriptions of the algorithms.

rfeIter is the basic algorithm while rfe wraps these operations inside of resampling. To avoid selection bias, it is better to use the function rfe than rfeIter.

See Also

rfeControl

Examples

data(BloodBrain)

x <- scale(bbbDescr[,-nearZeroVar(bbbDescr)])
x <- x[, -findCorrelation(cor(x), .8)]
x <- as.data.frame(x)

set.seed(1)
lmProfile <- rfe(x, logBBB,
                 sizes = c(2:25, 30, 35, 40, 45, 50, 55, 60, 65),
                 rfeControl = rfeControl(functions = lmFuncs, 
                                         number = 200))
set.seed(1)
lmProfile2 <- rfe(x, logBBB,
                 sizes = c(2:25, 30, 35, 40, 45, 50, 55, 60, 65),
                 rfeControl = rfeControl(functions = lmFuncs, 
                                         rerank = TRUE, 
                                         number = 200))

xyplot(lmProfile$results$RMSE + lmProfile2$results$RMSE +
       rfProfile$results$RMSE + rfProfile2$results$RMSE ~ 
       lmProfile$results$Variables, 
       type = c("g", "p", "l"), 
       auto.key = TRUE)

rfProfile <- rfe(x, logBBB,
                 sizes = c(2, 5, 10, 20),
                 rfeControl = rfeControl(functions = rfFuncs))

bagProfile <- rfe(x, logBBB,
                  sizes = c(2, 5, 10, 20),
                  rfeControl = rfeControl(functions = treebagFuncs))

set.seed(1)
svmProfile <- rfe(x, logBBB,
                  sizes = c(5, 20, 65),
                  rfeControl = rfeControl(functions = caretFuncs, 
                                          number = 200),
                  ## pass options to train()
                  method = "svmRadial",
                  fit = FALSE)

## classification with no resampling

data(mdrr)
mdrrDescr <- scale(mdrrDescr[,-nearZeroVar(mdrrDescr)])
mdrrDescr <- mdrrDescr[, -findCorrelation(cor(mdrrDescr), .8)]

set.seed(1)
inTrain <- createDataPartition(mdrrClass, p = .75, list = FALSE)[,1]

train <- mdrrDescr[ inTrain, ]
test  <- mdrrDescr[-inTrain, ]
trainClass <- mdrrClass[ inTrain]
testClass  <- mdrrClass[-inTrain]

preProc <- preProcess(train)
train <- predict(preProc, train)
test  <- predict(preProc, test)

nbProfile <- rfeIter(train, trainClass,
                     test,  testClass, 
                     sizes = c(1:10, 15, 30),
                     rfeControl = rfeControl(functions = nbFuncs))

splitUp <- split(nbProfile$pred, 
                 factor(nbProfile$pred$subset))
testResults <- lapply(splitUp, 
                      function(u) postResample(u$pred, u$obs))
Variables <- as.numeric(names(testResults))

testResults <- do.call("rbind", testResults)
testResults <- cbind(testResults, Variables)
plot(testResults[,3], testResults[,1])

#######################################
## Parallel Processing Example via MPI

## A function to emulate lapply in parallel
mpiClacs <- function(X, FUN, ...)
  {
    theDots <- list(...)
    parLapply(theDots$cl, X, FUN)
  }

library(snow)
cl <- makeCluster(5, "MPI")

set.seed(1)
lmProfile <- rfe(x, logBBB,
                 sizes = c(2:25, 30, 35, 40, 45, 50, 55, 60, 65),
                 rfeControl = rfeControl(functions = lmFuncs, 
                                         number = 200,
                                         workers = 5,
                                         computeFunction = mpiClacs,
                                         computeArgs = list(cl = cl)))

stopCluster(cl)

#######################################
## Parallel Processing Example via NWS
nwsClacs <- function(X, FUN, ...)
  {
    theDots <- list(...)
    eachElem(theDots$sObj,
             fun = FUN,
             elementArgs = list(X))
  }

library(nws)
sObj <- sleigh(workerCount = 5)

set.seed(1)
lmProfile <- rfe(x, logBBB,
                 sizes = c(2:25, 30, 35, 40, 45, 50, 55, 60, 65),
                 rfeControl = rfeControl(functions = lmFuncs, 
                                         number = 200,
                                         workers = 5,
                                         computeFunction = nwsClacs,
                                         computeArgs = list(sObj = sObj)))
close(sObj)