Create a Hybrid Sequential Feature Selector

Convenience function to create a SequentialSelector with common presets.

Usage

create_hsfs_selector(preset = "default", ...)

Arguments

preset

One of "default", "aggressive", "conservative", or "custom"

...

Additional arguments passed to SequentialSelector$new()

Value

A SequentialSelector object

Examples

if (FALSE) { # \dontrun{
if (requireNamespace("mlr3learners", quietly = TRUE) &&
    requireNamespace("ranger", quietly = TRUE) &&
    requireNamespace("xgboost", quietly = TRUE)) {
  # Default pipeline: variance -> ANOVA(5000) -> RF_importance(1000) -> AUC_filter
  selector <- create_hsfs_selector("default")

  # Aggressive reduction for very high-dimensional data
  selector <- create_hsfs_selector("aggressive")

  # Conservative - keep more features
  selector <- create_hsfs_selector("conservative")

  # Create GraphLearner
  learner <- selector$create_learner(model = "xgboost")
}
} # }