Skip to contents

Plot variable importance scores

Source: R/vip.R
plot.vi.Rd

Plot variable importance scores from a vi object using lightweight base R graphics (via the tinyplot package). This is the workhorse behind vip, which is a convenience wrapper that computes the scores and plots them in one call.

Usage

# S3 method for class 'vi'
plot(
  x,
  type = c("bar", "point", "boxplot", "violin"),
  num_features = 10L,
  horizontal = TRUE,
  all_permutations = FALSE,
  jitter = FALSE,
  include_type = FALSE,
  ...
)

Arguments

x

A vi object (i.e., the output of vi() or one of the vi_*() functions).

type

Character string specifying which type of plot to construct. The currently available options are described below.

  • type = "bar" (the default) constructs a bar chart of the variable importance scores.

  • type = "point" constructs a Cleveland dot plot of the variable importance scores.

  • type = "boxplot" constructs a boxplot of the raw permutation scores for each feature. This option can only be used for the permutation-based importance method with nsim > 1 and keep = TRUE; see vi_permute for details.

  • type = "violin" constructs a violin plot of the raw permutation scores for each feature. This option can only be used for the permutation-based importance method with nsim > 1 and keep = TRUE; see vi_permute for details.

num_features

Integer specifying the number of variable importance scores to plot. Default is 10.

horizontal

Logical indicating whether or not to plot the importance scores on the x-axis (TRUE). Default is TRUE.

all_permutations

Logical indicating whether or not to plot all permutation scores along with the average. Default is FALSE. (Only used for permutation scores when nsim > 1.)

jitter

Logical indicating whether or not to jitter the raw permutation scores. Default is FALSE. (Only used when all_permutations = TRUE.)

include_type

Logical indicating whether or not to include the type of variable importance computed in the axis label. Default is FALSE.

...

Additional graphical parameters passed on to tinyplot (e.g., col, fill, pch, cex, or main).

Value

Draws a plot as a side effect and invisibly returns the (trimmed and sorted) vi object being plotted.

Examples

# Fit a projection pursuit regression model
model <- ppr(mpg ~ ., data = mtcars, nterms = 1)

# Compute permutation-based variable importance scores
set.seed(825)  # for reproducibility
pfun <- function(object, newdata) predict(object, newdata = newdata)
vis <- vi(model, method = "permute", train = mtcars, target = "mpg",
          nsim = 10, metric = "rmse", pred_wrapper = pfun)

# Plot the results; additional arguments are passed on to
# `tinyplot::tinyplot()`
plot(vis)

plot(vis, type = "point", horizontal = FALSE, col = "forestgreen", cex = 2)

plot(vis, type = "boxplot", all_permutations = TRUE, jitter = TRUE,
     fill = "grey90")