Lattice functions for plotting resampling results of recursive feature selection
A set of lattice functions are provided to plot the resampled performance estimates (e.g. classification accuracy, RMSE) over different subset sizes.
## S3 method for class 'rfe' densityplot(x, data = NULL, metric = x$metric, ...)
x |
An object produced by |
data |
This argument is not used |
metric |
A character string specifying the single performance metric that will be plotted |
... |
arguments to pass to either
|
By default, only the resampling results for the optimal model are saved in
the rfe object. The function rfeControl can be used to
save all the results using the returnResamp argument.
If leave-one-out or out-of-bag resampling was specified, plots cannot be
produced (see the method argument of rfeControl)
A lattice plot object
Max Kuhn
## Not run:
library(mlbench)
n <- 100
p <- 40
sigma <- 1
set.seed(1)
sim <- mlbench.friedman1(n, sd = sigma)
x <- cbind(sim$x, matrix(rnorm(n * p), nrow = n))
y <- sim$y
colnames(x) <- paste("var", 1:ncol(x), sep = "")
normalization <- preProcess(x)
x <- predict(normalization, x)
x <- as.data.frame(x, stringsAsFactors = TRUE)
subsets <- c(10, 15, 20, 25)
ctrl <- rfeControl(
functions = lmFuncs,
method = "cv",
verbose = FALSE,
returnResamp = "all")
lmProfile <- rfe(x, y,
sizes = subsets,
rfeControl = ctrl)
xyplot(lmProfile)
stripplot(lmProfile)
histogram(lmProfile)
densityplot(lmProfile)
## End(Not run)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.