|
MODEL CONFIDENCE FOR NONLINEAR SYSTEMS
Wayne J. Dunstan* Robert R. Bitmead*
* Department of Mechanical & Aerospace Engineering,
University of California San Diego,
La Jolla CA 92093-0411, USA
This paper deals with defining measures of model closeness and establishing quantitative confidence bounds on nominal models. Confidence in a model is an indication of how uniquely identifiable the best fitting parameter values are from the data. These concepts are examined in both the linear and nonlinear regimes, with a practical example used to explore these propositions.
Keywords: Model Confidence, Nonlinear Modeling, Combustion Instability, System Identification
Session slot T-Th-A01: Identification of Nonlinear Systems II/Area code 3a : Modelling, Identification and Signal Processing
 | 