fit_P3_info_MCP: Information-adaptive minimax concave penalty (P3)

In fdb: Frequentist Dynamic Borrowing for Hybrid-Control Survival Trials

Information-adaptive minimax concave penalty (P3)

Description

Implements

p_\lambda(\delta) = MCP(\delta;\,\lambda/\widehat{SE}(\hat\delta_0),\,\gamma_{MCP}),

where MCP is the minimax concave penalty of Zhang (2010). Reduces bias when moderate population drift is present while retaining strong shrinkage near \delta = 0.

Usage

fit_P3_info_MCP(
  dat,
  xnames = NULL,
  lambda,
  gamma_mcp = 3,
  delta_bounds = DEFAULT_DELTA_BOUNDS,
  full_fit = NULL,
  robust = FALSE,
  eps = SMOOTH_EPS,
  n_grid_opt = DEFAULT_N_GRID_OPT,
  rho_mcp = DEFAULT_RHO_MCP
)

Arguments

dat	A data frame from simulate_hybrid_cox or conforming to its column conventions.
xnames	Character vector of covariate names. If NULL, automatically detected.
lambda	Penalty strength (\lambda \ge 0).
gamma_mcp	MCP shape parameter (> 1).
delta_bounds	Numeric vector of length 2 giving the optimization interval for \delta.
full_fit	Optional pre-computed full Cox fit (output of an internal helper). If NULL, the fit is computed.
robust	Use robust (Lin-Wei) Cox standard errors.
eps	Smoothing parameter for |\delta|_\varepsilon.
n_grid_opt	Number of grid points for the coarse search in the non-convex objective.
rho_mcp	MCP transition fraction in (0, 1); h = rho_mcp * gamma_mcp * lambda_eff. The software default 0.1 is not calibrated.

Details

MCP is smoothed at both the origin and the flat-tail transition by integrating the continuously differentiable slope described in the manuscript. First-stage quantities, including h, are held fixed.

Value

A list of estimates and inference quantities, including the effective MCP lambda_eff and the raw curvature pen_curv_raw.

References

Zhang, C.-H. (2010). Nearly unbiased variable selection under minimax concave penalty. The Annals of Statistics 38(2), 894-942.

fdb documentation built on Oct. 4, 2026, 5:07 p.m.