The International Institute for Applied Systems Analysis (IIASA) cordially invites you to a research talk on October 20th, with
 
Florian Huber
(University of Salzburg)
on
“Bayesian Modeling of Time-varying Parameters Using Regression Trees”
 
Abstract: In light of widespread evidence of parameter instability in macro-Economic models, many time-varying parameter (TVP) models have been proposed. This paper proposes a nonparametric TVP-VAR model using Bayesian Additive Regression Trees (BART). The novelty of this model arises from the law of motion driving the parameters being treated nonparametrically. This leads to great flexibility in the nature and extent of parameter change, both in the conditional mean and in the conditional variance. In contrast to other nonparametric and machine learning methods that are black box, inference using our model is straightforward because, in treating the parameters rather than the variables nonparametrically, the model remains conditionally linear in the mean. Parsimony is achieved through adopting nonparametric factor structures and use of shrinkage priors. In an application to US macroeconomic data, we illustrate the use of our model in tracking both the evolving nature of the Phillips curve and how the effects of business cycle shocks on inflationary measures vary nonlinearly with movements in uncertainty.
 
Date 

Thursday, October 20th
 
Time 

2.00 to 3.30 pm (CET)
 
Location: 

Wodak room

International Institute for Applied Systems Analysis (IIASA)

Schloßplatz 1, 2361

Laxenburg, Austria
 

For administrative reasons we would kindly ask you to register here if you would like to attend this event. 

 
With my best regards,

Tamás Krisztin