Wolfram Library Archive


Courseware Demos MathSource Technical Notes
All Collections Articles Books Conference Proceedings
Title Downloads

Data Driven Model Estimation: A System Identification Package for Predicting Time-Series and to Enhance Dynamical Models from Data
Authors

Jonas Sjöberg
Organization: Chalmers University of Technology
Håkan Hjalmarsson
Organization: Royal Institute of Technology
Conference

2006 Wolfram Technology Conference
Conference location

Champaign IL
Description

This talk describes a system identification package that is under development. Previews of algorithms and supported model structures are given in connection with examples. The package supports algorithms for identification of linear and nonlinear models of dynamical systems that can be used for simulation, prediction, and for control design. The symbolic features of Mathematica are used so that a large variety of model structures can be handled. Before the numerical estimation procedure, the symbolic definition of the model structure is simplified and modified before the numerical computations of the parameters. In this way a larger set of model structures can be supported, at the same time as the computational speed is improved.
Subjects

*Applied Mathematics > Information Theory
*Engineering > Signal Processing
Keywords

time series, model estimation, data
Downloads Download Wolfram CDF Player

Download
DataDrivenModelEstimation.nb (3.3 MB) - Mathematica Notebook [for Mathematica 5.2]