Bayesian Parameter Estimation for Stochastified Compartmental Models with Partial Data Based on Mixed Graphical Models and Applications to the Karlsburg Model for the Glucose-Insulin Metabolism
| Autor: | Eckhard Salzsieder, Ulrich G. Oppel, Gabriele Bleckert |
|---|---|
| Rok vydání: | 2000 |
| Předmět: |
Discretization
Computer science Differential equation business.industry System identification Machine learning computer.software_genre Bayesian parameter estimation Discrete system Set (abstract data type) Stochastic differential equation Applied mathematics Artificial intelligence Graphical model business computer |
| Zdroj: | IFAC Proceedings Volumes. 33:297-302 |
| ISSN: | 1474-6670 |
| DOI: | 10.1016/s1474-6670(17)35531-3 |
| Popis: | We transform a compartmental model given by a set of deterministic differential equations and a set of discrete system parameters into a system of stochastic differential equations and obtain a mixed graphical model by time discretization. This mixed graphical model allows for Bayesian parameter estimation by introducing partial knowledge about the system. This improved knowledge then may be used for prediction and, hence, optimization of therapy. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|