Summarize selected CI tuning fits — ci_fit_summary

Builds a report-ready data table for selected tune_ci_tree() or tune_ci_forest() settings. The helper combines wide tuning metrics with mean root objective values from validation folds when they are available.

Usage

ci_fit_summary_table(
  x,
  selected = ci_select_best(x),
  metrics = c("train_gain", "validation_gain", "train_relative_gain",
    "relative_validation_gain"),
  include_percent = TRUE,
  ...
)

Arguments

x: A tuning object returned by tune_ci_tree() or tune_ci_forest().
selected: Optional selected tuning rows. Defaults to ci_select_best() using the tuning object's selection metric.
metrics: Metrics to include before widening. Training metrics such as "train_gain" and "train_relative_gain" are allowed when the tuning object contains training diagnostics.
include_percent: Logical; add percent gain columns relative to the absolute mean root objective when possible.
...: Reserved for future extensions.

Value

A data.table.