Meinshausen's Hierarchical Testing with Sequential Rejection Principle — meinshausen_hierarchical

Implementation of Meinshausen's (2008) hierarchical testing of variable importance enhanced with the sequential rejection principle of Goeman and Solari (2010). This approach provides a unified framework for hierarchical testing that maintains FWER control while leveraging the hierarchical structure for increased power.

Usage

meinshausen_hierarchical_test(
  node_dat,
  node_tracker,
  alpha = 0.05,
  method = "simes",
  use_sequential = TRUE
)

Arguments

node_dat: Data.table containing node-level test results with columns: nodenum, p, depth, nodesize, and optionally parent, testable
node_tracker: Node tracking object containing the hierarchical structure
alpha: Global Type I error rate (default: 0.05)
method: Method for combining p-values ("simes", "fisher", "bonferroni")
use_sequential: Logical, whether to use sequential rejection principle

Value

Updated node_dat with Meinshausen hierarchical testing results

References

Meinshausen, N. (2008). Hierarchical testing of variable importance. Biometrika 95, 265-278.

Goeman, J. J., & Solari, A. (2010). The sequential rejection principle of familywise error control. Annals of Statistics 38, 3782-3810.