Estimate oracle prediction function using DR gradient boosting

Estimate oracle prediction function using DR gradient boosting

Usage

generate_oracle_predictions_boost(
  time,
  event,
  X,
  X_holdout,
  nuisance_preds,
  restriction_time,
  approx_times,
  V = 5,
  indx,
  tuning = FALSE,
  subsample_n = length(time),
  boosting_params = list(mstop = c(100), nu = c(0.1), sigma = c(0.01), learner =
    c("glm"))
)

Arguments

time

n x 1 numeric vector of observed follow-up times. If there is censoring, these are the minimum of the event and censoring times.

event

n x 1 numeric vector of status indicators of whether an event was observed.

X

n x p data.frame of observed covariate values

X_holdout

m x p data.frame of new observed covariate values at which to obtain m predictions for the estimated algorithm. Must have the same names and structure as X.

nuisance_preds

Named list of conditional survival function predictions with elements "S_hat", "S_hat_train", "G_hat", and "G_hat_train". This should match the output of conditional_surv_generator.

restriction_time

Maximum follow-up time for calculation of the C-index. Essentially, this time should be chosen such that the conditional survival function is identified at this time for all covariate values X present in the data. Choosing the restriction time such that roughly 10% of individuals remain at-risk at that time has been shown to work reasonably well in simulations.

approx_times

Numeric vector of length J2 giving times at which to approximate C-index integral.

V

Number of cross-validation folds for selection of tuning parameters

indx

Numeric index of column(s) of X to be removed, i.e., not used in the oracle prediction function.

tuning

Logical, whether or not to use cross-validation to select tuning parameters

subsample_n

Number of samples to use for boosting procedure. Using a subsample of the full sample can greatly reduce runtime

boosting_params

Named list of parameter values for the boosting procedure. Elements of this list include mstop (number of boosting iterations), nu (learning rate), sigma (smoothness parameter for sigmoid approximation, with smaller meaning less smoothing), and learner (base learner, can take values "glm", "gam", or "tree")

Value

A list containing elements f0_hat and f0_hat_train, the estimated oracle prediction functions for X_holdout and X, respectively.

See also

boost_c_index