Example 32 — Nonparametric Support Points

Script: examples/32_nonparametric_support_points.py

This example demonstrates nonparametric population estimation on a synthetic one-compartment oral PK dataset with a single ETA on clearance. The synthetic dataset keeps runtime short while still producing a non-degenerate support-point distribution.

What it shows

  • NONPARAMETRIC estimation with a deterministic synthetic dataset.

  • A support-point approximation to the empirical ETA distribution.

  • The highest-weight support points and empirical ETA moments.

Model

The example uses:

  • ADVAN2 / TRANS2

  • fixed structural parameters KA and V

  • one random effect on CL

  • proportional residual error

This narrower setup is deliberate: it is much easier to obtain an informative, fast support-point example with one ETA than with a heavier multi-ETA real-data workflow.

Notes

  • The printed converged flag comes from the base parametric fit used to seed the support points. The support-point distribution itself is the important artifact in this example.

  • The example is intended as a runnable illustration, not an external scientific reference dataset.