Fast Weibull Regression with General Censoring (C++ Backend)

Weibull Accelerated Failure Time model fitting extended to left-, right-, and interval-censored responses (TODO-3 in interval_censored_survival_response.md). Zero-regression by construction: exact/right-censored-only input uses the same likelihood contributions as the corresponding survival::Surv() response.

Usage

fast_weibull_regression_general_cpp(
  X,
  y,
  y_L,
  y_R,
  warm_start_params = NULL,
  smart_cold_start = TRUE,
  estimate_only = FALSE,
  maxit = 100L,
  tol = 1e-08,
  fixed_idx = NULL,
  fixed_values = NULL,
  optimization_alg = "lbfgs",
  warm_start_fisher_info = NULL
)

Arguments

X

A numeric matrix of predictors.

y

Exact survival times, NA for censored subjects.

y_L

Censored-interval lower bounds, NA for exact subjects; 0 for left-censored.

y_R

Censored-interval upper bounds, NA for exact subjects; Inf for right-censored.

warm_start_params

Optional starting values for coefficients.

smart_cold_start

Logical. If TRUE, use an initial OLS-based guess.

estimate_only

Logical. If TRUE, do not compute variance-covariance.

maxit

Maximum number of iterations.

tol

Convergence tolerance.

fixed_idx

Optional indices of fixed parameters.

fixed_values

Optional values for fixed parameters.

optimization_alg

Optimization algorithm.

warm_start_fisher_info

Optional initial Fisher Information matrix.

Value

A list containing the following components:

coefficients

A numeric vector of the estimated Weibull regression coefficients, including the intercept.

log_sigma

The logarithm of the scale parameter from the Weibull distribution.

vcov

The variance-covariance matrix of the estimated coefficients.

neg_ll

The negative log-likelihood at the final iteration.

converged

A logical value indicating whether the algorithm converged.

A list containing coefficients, log_sigma, and convergence status.