Artikel
Lizenziert
Nicht lizenziert
Erfordert eine Authentifizierung
Learning under signal-to-noise ratio uncertainty
-
Alex Ilek
Veröffentlicht/Copyright:
14. Februar 2013
Abstract
The paper presents an alternative real time adaptive learning algorithm in the presence of signal-to-noise ratio uncertainty. The main innovation of this algorithm is that it uses a gain which is determined within the model: it continuously depends on the extent of misevaluation of parameters embedded in the forecast error. We show that in the presence of signal-to-noise ratio misevaluation, the usage of the proposed learning algorithm is a significant improvement on the Kalman Filter learning algorithm. In a full information case, the Kalman Filter learning algorithm is still the optimal tool.
Keywords: adaptive learning; endogenous gain; Kalman Filter; parameter misevaluation index; signal-to-noise ratio
Published Online: 2013-02-14
©2013 by Walter de Gruyter Berlin Boston
Sie haben derzeit keinen Zugang zu diesem Inhalt.
Sie haben derzeit keinen Zugang zu diesem Inhalt.
Artikel in diesem Heft
- Forecast uncertainty and the Bank of England’s interest rate decisions
- A Bayesian approach for capturing daily heterogeneity in intra-daily durations time series
- Learning under signal-to-noise ratio uncertainty
- Using transfer entropy to measure information flows between financial markets
- Computational aspects of portfolio risk estimation in volatile markets: a survey
Schlagwörter für diesen Artikel
adaptive learning;
endogenous gain;
Kalman Filter;
parameter misevaluation index;
signal-to-noise ratio
Artikel in diesem Heft
- Forecast uncertainty and the Bank of England’s interest rate decisions
- A Bayesian approach for capturing daily heterogeneity in intra-daily durations time series
- Learning under signal-to-noise ratio uncertainty
- Using transfer entropy to measure information flows between financial markets
- Computational aspects of portfolio risk estimation in volatile markets: a survey
