Skip to contents

Provides a narrative interpretation of policy tree stability results suitable for inclusion in scientific manuscripts. The interpretation acknowledges the inherent instability of decision trees while focusing on robust patterns that emerge across iterations.

Usage

margot_interpret_stability(
  object,
  model_name,
  depth = 2,
  stability_threshold = 0.7,
  format = c("text", "technical"),
  decimal_places = 1,
  include_theory = TRUE,
  label_mapping = NULL
)

Arguments

object

Object of class "margot_stability_policy_tree"

model_name

Model name to interpret

depth

Tree depth to interpret (1, 2, or "both")

stability_threshold

Minimum frequency to consider a split "stable" (default 0.7)

format

Output format: "text" for narrative prose or "technical" for detailed statistics

decimal_places

Number of decimal places for statistics (default 1)

include_theory

Logical: Include theoretical context about tree instability (default TRUE)

label_mapping

Optional named list mapping variable names to labels. If NULL, uses automatic transformation via transform_var_name()

Value

Character string containing the interpretation