Abstract
Parameter Identifiability Analysis (IA) of large scale
dynamical models is a technically tedious process. The realization
of this task can benefit from a platform assisting the
easier task of determining small identifiable parameter subsets.
Then, these subsets get enlarged on an iterative basis until
a maximal identifiable parameter set is obtained. Moreover,
the employment of mutual combinations of many candidates
of mathematical tools and algorithmic variants for the underlying
interrelated computational subtasks within Parameter
Estimation (PE) needs to be examined. This work demonstrates
the developer-oriented software DecTrees employed
for establishing such a platform. It provides a compact generic
implementation of Decision Trees (DTs). The main elements are
represented via context-free C++ components describing nodes
and conditioned edges. By extending these white-boxes with a
context, meaningful applicative decision systems are established.
This is the case with the software InteractiveMenus useful
for simplifying the configuration of complicated computational
tasks. The advantages are emphasized with a realistic IA
application. The software is open-source and is available under
https://github.com/AtiyahElsheikh/DecTrees.
Originalsprache | Englisch |
---|---|
Titel | The 8th EUROSIM Congress on Modelling and Simulation, Cardiff, Wales, Septemeber 2013 |
Seiten | 300-305 |
Seitenumfang | 6 |
Publikationsstatus | Veröffentlicht - 2013 |
Research Field
- Ehemaliges Research Field - Energy
Schlagwörter
- decision trees; identifiability analysis; dynamical systems;