Compute held-out concentration-index validation gain — ci_tree_validation

Applies a fitted greedy CI tree to validation data, assigns each validation observation to a terminal node, and computes the reduction from root concentration-index impurity to weighted within-terminal-node impurity.

A larger value means that the fitted partition leaves less within-node socioeconomic inequality in the held-out data.

Usage

ci_tree_validation_gain(
  fit,
  new_data,
  rank_name,
  outcome_name,
  weights = NULL,
  type = c("CI", "CIg", "CIc", "L"),
  root_impurity = NULL
)

Arguments

fit: A fitted ci_tree or other object inheriting from party.
new_data: Validation data.
rank_name: Name of the socioeconomic ranking column.
outcome_name: Name of the outcome column.
weights: Optional non-negative validation weights.
type: One of "CI", "CIg", "CIc", or "L" selecting the inequality index used for validation scoring. "L" uses observed socioeconomic levels in the first response column rather than fractional ranks.
root_impurity: Optional pre-computed root impurity for the validation sample under the selected concentration-index criterion. If NULL, the root impurity is computed from new_data.

Value

A numeric validation gain.

Examples

toy_data <- data.frame(
  rank = c(10, 20, 30, 40, 50, 60),
  outcome = c(1, 0, 1, 0, 1, 1),
  income = c(2, 4, 6, 8, 10, 12)
)
fit <- ci_tree(
  cbind(rank, outcome) ~ income,
  data = toy_data,
  rank_name = "rank",
  outcome_name = "outcome",
  control = ci_tree_control(minsplit = 1, minbucket = 1, maxdepth = 1)
)
ci_tree_validation_gain(fit, toy_data, "rank", "outcome")
#> [1] 0.08333333