Solving ODEs in the wild: Scalable pharmacometrics with Stan
| Main Author: | Weber, Sebastian |
|---|---|
| Format: | Proceeding Journal |
| Terbitan: |
, 2018
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/1465996 |
Daftar Isi:
- Pharmacometric modeling involve nonlinear hierarchical models which are most naturally expressed as ordinary differential equations. These class of models lead to a number of challenges which complicate a practical modeling work-flow in Stan. At the example of the public Warfarin data-set I will present solutions to these. I will first illustrate the concept of forcing functions and how to use these efficiently in Stan. Moreover, I will show how the hierarchical structure of the problem is most efficiently taken advantage of to parallelize the model computations which expedites the model evaluation time and allows to scale Stan’s performance to large data-sets given sufficient computing resources.
- Code and data available at github.com/stan-dev/stancon_talks