Competing Risks with Relative Entropy Integration using XGBoost and Composite Failure Time Prior

This function implements the Relative Entropy (RE) framework for discrete-time competing risks models where the prior model is specified in a composite failure time formulation and the local model is fit using XGBoost.

Usage

CompRiskRE_XGBoost_FT(
  prior_model,
  train_data,
  test_data,
  eta,
  xgb_params = list(objective = "multi:softprob", eval_metric = "mlogloss", num_class =
    3, max.depth = 4),
  nrounds = 100,
  maxiter = 1,
  eps = 1e-06
)

Arguments

prior_model

A fitted prior multinomial model.

train_data

A data.frame containing training data.

test_data

A data.frame containing test data.

eta

Numeric vector of RE regularization parameters.

xgb_params

A list of XGBoost parameters..

nrounds

Integer, number of boosting rounds for XGBoost (default: 100).

maxiter

Integer, maximum number of RE updates in priorFTKL_XGBoost (default: 1).

eps

Numeric tolerance for convergence in priorFTKL_XGBoost (default: 1e-6).

Value

A list with two components:

models

A named list of fitted XGBoost models for each supplied eta.

PD

A numeric vector of predictive deviance values on the test data.

eta

The vector of eta values.