Best Subset Variable Selection with Blocked CV

Evaluates all 2^p combinations of predictor variables using blocked CV. Exhaustive search — only feasible for small p (< 15).

Usage

borg_best_subset(
  data,
  target,
  predictors,
  folds,
  metric = c("rmse", "mae", "rsq"),
  fit_fun = stats::lm,
  max_vars = NULL,
  verbose = FALSE
)

Arguments

data

Data frame.

target

Character. Response variable.

predictors

Character vector. Candidate predictors.

folds

A borg_cv object or fold list.

metric

Character. Default: "rmse".

fit_fun

Function. Default: lm.

max_vars

Integer. Maximum variables in a subset. Default: all.

verbose

Logical. Default: FALSE.

Value

A data frame with class "borg_bss": variables, n_vars, metric_value, rank.