Cross-Validation of Controlled Dynamic Models: Bayesian Approach
| Authors: | Karny Miroslav, UTIA, AV CR, Czech Republic
Nedoma Petr, UTIA, AV CR, Czech Republic
Smidl Vaclav, UTIA, AV CR, Czech Republic |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Model Validation Techniques |
|---|
| Keywords: | validation, estimation theory, recursive estimation |
|---|
Abstract
The best test of quality of an estimated model is its implementation in real application.However, the use of a bad model is typically too costly.Therefore, model validation is considered as an obligatory step in model learning, andextensive theory has been developed within statistical community.However, the available rules deal almostexclusively with independent data samples. Consequently, they aresubstantially disqualified for validation of dynamicmodels.This paper approaches the problem using Bayesian formulation and solution.An algorithm for validation of models estimated withinpractically important exponential family is presented. Performance of thealgorithms is illustrated on simulated example.
