Skip to contents

Plot Outer-Fold Coefficient Importance and Selection Frequency

Source: R/compare_censoring_weights.R
plot_outer_importance.Rd

Visualizes predictor coefficient distributions and selection frequencies across outer resampling folds. Accepts output from [tidy_outer_coefficients()], [score_outer()], or [summarize_outer_importance()].

Usage

plot_outer_importance(
  x,
  top_n = NULL,
  nonzero_only = FALSE,
  type = c("boxplot", "frequency"),
  label_map = NULL,
  style = NULL
)

Arguments

x

A tibble from [tidy_outer_coefficients()], [score_outer()], or [summarize_outer_importance()].

top_n

Optional integer restricting the plot to the top `top_n` predictors ranked by selection frequency and absolute median coefficient.

nonzero_only

Logical (default `FALSE`). When `TRUE`, drops zero coefficient estimates before drawing boxplots.

type

Plot type: `"boxplot"` (default; horizontal boxplot or range intervals of coefficients annotated with selection percentage) or `"frequency"` (bar/lollipop chart of selection frequency).

label_map

Optional label mapping passed to [resolve_cbe_labels()].

style

Label style (`"short"` or `"full"`) when resolving `label_map`.

Value

A `ggplot` object.

See also

[summarize_outer_importance()], [tidy_outer_coefficients()], [score_outer()]