Abstract
Parameter estimation in general state space models is not trivial as the likelihood is unknown. We propose a recursive estimator for general state space models, and show that the estimates converge to the true parameters with probability one. The estimates are also asymptotically Cramer-Rao efficient. The proposed estimator is easy to implement as it only relies on non-linear filtering. This makes the framework flexible as it is easy to tune the implementation to achieve computational efficiency. This is done by using the approximation of the score function derived from the theory on Iterative Filtering as a building block within the recursive maximum likelihood estimator.
| Original language | English |
|---|---|
| Title of host publication | IFAC-PapersOnLine (System Identification, Volume 16) |
| Publisher | Elsevier |
| Pages | 1785-1790 |
| Number of pages | 6 |
| ISBN (Print) | 978-3-902823-06-9 (online) |
| DOIs | |
| Publication status | Published - 2012 |
| Event | 16th IFAC Symposium on System Identification - Brussels, Belgium Duration: 2012 Jul 11 → 2012 Jul 13 |
Conference
| Conference | 16th IFAC Symposium on System Identification |
|---|---|
| Country/Territory | Belgium |
| City | Brussels |
| Period | 2012/07/11 → 2012/07/13 |
Bibliographical note
The paper is accessible (free of charge) to the public . To download the paper an account at IFAC is needed. The web browser Internet Explorer is recommended.Subject classification (UKÄ)
- Probability Theory and Statistics
Free keywords
- Recursive estimation
- maximum likelihood estimator
- filtering techniques
- stochastic approximation
- iterative methods