Tore Selland Kleppe

Associate Professor of statistics

Interests::

Computational statistics; Dynamic latent variable models; Econometrics

Adress:

Department of Mathematics
University of Bergen
Johannes Bruns gate 12
N-5008 Bergen
Norway
Phone: (+47) 5558 4871
eMail: tore.kleppe[at]math.uib.no

Publications

- Building and fitting non-Gaussian latent variable models via the moment generating function, (2008) (with Hans Skaug)
Scandinavian Journal of Statistics 35 (4) pages 664-676

-Simulated Maximum Likelihood Estimation of Continuous Time Stochastic Volatility Model (2010) (with Jun Yu and Hans Skaug)
Advances in Econometrics: Maximum Simulated Likelihood, 26

-Asymptotic bias of the hazard probability model under model miss-specification (2010) (with Hans Skaug and Hiroshi Okamura)
The Journal of Cetacean Research and Management

- Fitting general stochastic volatility models using Laplace accelerated sequential importance sampling (2012) (with Hans Skaug)
Computational Statistics and Data Analysis 56 (11) pages 3106-3119

- Efficient high-dimensional importance sampling in mixture frameworks (2013) (with Roman Liesenfeld)
Forthcoming Computational Statistics and Data Analysis.

Working Papers

-Simulated Maximum Likelihood Estimation for Latent Diffusion Models (2011) (with Jun Yu and Hans Skaug)
Submitted.

- Bandwidth selection in pre-smoothed particle filters (updated December 2012) (with Hans Skaug) Supplementary material
Submitted.

- Estimating the competitive storage model: a simulated likelihood approach (February 2013) (with Atle Øglend)
Submitted.

Software

- localgauss; an R-package implementing computational and graphical routines for local Gaussian parameters

Master Thesis

My master thesis; "Numerical Path Integration for Lévy Driven Stochastic Differential
Equations" may be found here. My supervisor at the NTNU was Arvid Næss.