Here we summarize some results that are further developed in Nava et al. (2013). We pro-pose a method to construct strictly stationary Markovian models with fixed invariant distributions. Of particular interest are those models with invariant distributions belonging to the class of Generalized Inverse Gaussian (GIG) distributions family. The construction we propose is based on a Poisson trans-form which controls the dependence structure in the model. In particular, it allows to fully control the underlying transition probabilities that, an appealing feature, is then incorporated within standard es-timation methods. A Bayesian estimate via a Gibbs sampler algorithm, based on the slice method, is proposed and implemented.
On Stationary Markov Models: a Poisson-driven approach
PRUENSTER, IGOR
2013
Abstract
Here we summarize some results that are further developed in Nava et al. (2013). We pro-pose a method to construct strictly stationary Markovian models with fixed invariant distributions. Of particular interest are those models with invariant distributions belonging to the class of Generalized Inverse Gaussian (GIG) distributions family. The construction we propose is based on a Poisson trans-form which controls the dependence structure in the model. In particular, it allows to fully control the underlying transition probabilities that, an appealing feature, is then incorporated within standard es-timation methods. A Bayesian estimate via a Gibbs sampler algorithm, based on the slice method, is proposed and implemented.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.