Overview
Gregor Kastner is a statistician (WU Vienna University of Economics and Business; later University of Klagenfurt), working on Bayesian computation for time series — efficient MCMC for stochastic volatility, large VARs, and shrinkage.
Key Contributions / Features
- ASIS for stochastic volatility (Kastner-Frühwirth-Schnatter 2014): interweaving centered and non-centered parameterizations for efficient SV sampling; author of the
stochvol R package.
- Work on multivariate/factor stochastic volatility, sparse Bayesian time-varying-parameter models, and scalable Bayesian VARs.
Related