Endpoint alignment is public scientific metadata. It controls how a small fixed-dimensional trajectory summary is built and how new trajectories are generated; it is not a PK or PD model.
Usage
pmx_endpoint(
dvid = NULL,
alignment,
transform = c("auto", "log", "identity"),
shape,
units = NULL,
grid = NULL,
cmt = NULL,
subject_sd = 0.2,
residual_sd = 0.08,
censoring = NULL
)Arguments
- dvid
Public DVID value for this endpoint, or
NULLwhen no DVID column is used.- alignment
One of
"dose_relative","study_time","occasion", or"hybrid".- transform
One of
"log","identity", or"auto"."auto"uses only the public DV bounds: a nonnegative domain uses an offset log scale.- shape
One of
"occasion"or"global"; this is a broad public shape expectation, not a fitted structural model.- units
Optional public unit label.
- grid
Optional strictly increasing public grid on the declared clock. This is a discretization basis, not a sampling schedule. When omitted, a generic basis is constructed from public bounds and contribution limits; sampling-cell occupancy is then learned from the fitted data.
- cmt
Optional public observation compartment value.
- subject_sd, residual_sd
Public generation variability multipliers.
- censoring
Optional public list with
left,right, or a two-valueinterval. Source-dependent censoring frequencies are separately private.