Numerically Accelerated Importance Sampling for Nonlinear Non-Gaussian State Space Models
| Autor: | Siem Jan Koopman, Marcel Scharth, Andre Lucas |
|---|---|
| Přispěvatelé: | Econometrics and Operations Research, Finance, A-LAB, Tinbergen Institute, Econometrics and Data Science |
| Jazyk: | angličtina |
| Rok vydání: | 2015 |
| Předmět: |
Statistics and Probability
Economics and Econometrics Mathematical optimization Gaussian Monte Carlo method Efficient importance sampling jel:C22 Simulation smoothing Stochastic volatility model symbols.namesake Simulated maximum likelihood SDG 7 - Affordable and Clean Energy Mathematics Control variables Stochastic volatility State vector State space models importance sampling simulated maximum likelihood stochastic volatility stochastic copula stochastic conditional duration Kalman filter jel:C15 Numerical integration symbols Statistics Probability and Uncertainty Likelihood function Social Sciences (miscellaneous) Importance sampling |
| Zdroj: | Koopman, S J, Lucas, A & Scharth, M 2015, ' Numerically Accelerated Importance Sampling for Nonlinear Non-Gaussian State Space Models ', Journal of Business and Economic Statistics, vol. 33, no. 1, pp. 114-127 . https://doi.org/10.1080/07350015.2014.925807 Koopman, S J, Lucas, A & Scharth, M 2015, ' Numerically Accelerated Importance Sampling for Nonlinear Non-Gaussian State-Space Models ', Journal of Business and Economic Statistics, vol. 33, no. 1, pp. 114-127 . https://doi.org/10.1080/07350015.2014.925807 Journal of Business and Economic Statistics, 33(1), 114-127. American Statistical Association |
| ISSN: | 0735-0015 |
| Popis: | This paper led to a publication in the 'Journal of Business & Economic Statistics' , 2015, 33 (1), 114-127. We introduce a new efficient importance sampler for nonlinear non-Gaussian state space models. We propose a general and efficient likelihood evaluation method for this class of models via the combination of numerical and Monte Carlo integration methods. Our methodology explores the idea that only a small part of the likelihood evaluation problem requires simulation. We refer to our new method as numerically accelerated importance sampling. The method is computationally and numerically efficient, facilitates parameter estimation for models with high-dimensional state vectors, and overcomes a bias-variance trade-off encountered by other sampling methods. An elaborate simulation study and an empirical application for U.S. stock returns reveal large efficiency gains for a range of models used in financial econometrics. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|