Skip to contents

Predict mechanistic bank trajectories at required observation times

Source: R/bank.R
mbmm_predict_bank.Rd

Applies a user-supplied mechanistic simulator to every particle or multitarget bundle row in a reference bank.

Usage

mbmm_predict_bank(
  bank,
  simulate_fn,
  times = NULL,
  times_by_target = NULL,
  donor_col = "id",
  bundle_col = "bundle_id",
  failure = c("error", "drop")
)

Arguments

bank

Reference-bank data.frame.

simulate_fn

Function with signature simulate_fn(theta_row, times) returning one numeric prediction per requested time.

times

Numeric vector of evaluation times for single-target banks.

times_by_target

Named list of time vectors for multitarget banks.

donor_col

Name of the donor ID column.

bundle_col

Name of the bundle ID column.

failure

Either "error" or "drop". With "error", any failed or invalid simulation stops. With "drop", failed bank rows are omitted and reported in attributes.

Value

Long-format bank predictions containing donor ID, bundle ID, optional target, time, and prediction.