Helper function to predict from a super learner fit on the hazard scale to obtain a prediction on the density scale

predict_density(
  sl_fit_conditional,
  sl_fit_marginal,
  how_predict,
  all_mediator_values,
  a_val,
  validC,
  M1,
  M2,
  valid_n,
  stratify
)

Arguments

sl_fit_conditional

A fitted SuperLearner object for the M1 | M2, C, A regression

sl_fit_marginal

A fitted SuperLearner object for the M2 | C, A regression

how_predict

Needed to figure out how to predict from model. Possible choices are "SuperLearner", "single_algo", and "glm"

all_mediator_values

data.frame of all observed mediator values

a_val

Relevant treatment level

validC

Confounders in the validation set

M1

All M1 values

M2

All M2 values

valid_n

Number of obs in validation set

stratify

Were sl_fit's stratified?