SurvivalShrinkageBCF: SurvivalShrinkageBCF (Shrinkage Bayesian Causal Forest for survival data)

Description

Fits a survival version of a Bayesian Causal Forest (BCF) under an accelerated failure time (AFT) model, combining Dirichlet splitting priors with global–local shrinkage.

Usage

SurvivalShrinkageBCF(
  time,
  status,
  X_train,
  treatment,
  timescale = "time",
  propensity = NULL,
  a_dir = 0.5,
  b_dir = 1,
  number_of_trees_control = 200,
  number_of_trees_treat = 50,
  power_control = 2,
  base_control = 0.95,
  power_treat = 3,
  base_treat = 0.25,
  N_post = 1000,
  N_burn = 1000,
  verbose = TRUE,
  ...
)

Value

An object of class CausalShrinkageForest, containing posterior mean predictions, posterior samples (if stored), and estimated heterogeneous treatment effects. See CausalShrinkageForest for full details of returned components.

Arguments

time: Outcome vector of right-censored (non-negative) survival times.
status: Event indicator (1 = event, 0 = censored).
X_train: Design matrix for training data.
treatment: Treatment indicator (0/1) for training data.
timescale: Either "time" (log-transform internally) or "log" (already log-transformed).
propensity: Optional vector of propensity scores. If provided, it is appended to the control forest design matrix.
a_dir: First shape parameter of the Beta prior controlling the sparsity level in the Dirichlet splitting rule.
b_dir: Second shape parameter of the Beta prior controlling the sparsity level in the Dirichlet splitting rule.
number_of_trees_control: Number of trees in the control forest. Default is 200.
number_of_trees_treat: Number of trees in the treatment forest. Default is 50.
power_control, base_control: Tree-structure prior parameters for the control forest.
power_treat, base_treat: Tree-structure prior parameters for the treatment forest.
N_post: Number of posterior samples to store.
N_burn: Number of burn-in iterations.
verbose: Logical; print sampling progress.
...: Additional arguments passed to CausalShrinkageForest to override default hyperparameters.

Details

This wrapper extends SurvivalBCF by incorporating Dirichlet sparsity in both the prognostic (control) and treatment forests, while applying additional shrinkage to the control forest via a half-Cauchy prior.

The SurvivalShrinkageBCF model decomposes the outcome as $$ \log T = \mu(x) + a \cdot \tau(x) + \varepsilon, $$ where $\mu(x)$ represents the prognostic (control) component and $\tau(x)$ the heterogeneous treatment effect.

In contrast to SurvivalBCF, this function:

Applies a Dirichlet splitting prior to both forests, inducing structural sparsity in variable selection.
Combines Dirichlet sparsity with additional half-Cauchy shrinkage in the control forest.

The Dirichlet prior follows the sparse splitting framework of Linero (2018), where splitting probabilities are governed by a Beta–Dirichlet hierarchy. The sparsity level is controlled by a_dir and b_dir.

Survival outcomes are modeled using an AFT formulation with right-censoring handled via data augmentation.

References