[ABE-L] Large Bayesian Additive Vector Autoregressive Tree Models

[Hedibert Lopes]

Seminario Intersede de Estadística CIMAT
A toda la comunidad CIMAT - DEMAT

Cordialmente, los invitamos a una nueva sesión del Seminario Intersede de
Estadística. Los datos de la sesión del seminario están a continuación:

Fecha y hora: Miércoles 19 de marzo, 12:30-13:30 hrs. (GMT-6).
Sede Monterrey: Aula 1 B10

Sede Guanajuato: Diego Bricio (G101).

Resto de las sedes: Evento Virtual

ID de reunión: 946 4979 3652
Código de acceso: 502470
URL:  https://zoom.us/j/94649793652?pwd=JtYRKvqgGq5u3JmN1CLXfTKnxdUIp7.1

Ponente: Dr. Hedibert Lopes (INSPER)

Título: Large Bayesian Additive Vector Autoregressive Tree Models


Resumen: Vector autoregression (VAR) models are crucial for forecasting and
analyzing macroeconomic variables, serving as a fundamental tool for
applied macroeconomists. Recent literature has explored nonparametric
approaches, such as Bayesian additive regression trees (BART), which allow
for flexibility without strong parametric assumptions. However, existing
frameworks like the one proposed by Huber and Rossini (2022) do not
adequately accommodate high-dimensional data or time dependency in the
prior construction. This study enhances the literature by extending
previous work to enable high-dimensional data analysis and variable
selection through a sparsity-inducing Dirichlet hyperprior, as in Linero
(2018) on the regression tree’s splitting probabilities, while also
proposing a prior that decreases the probability of splitting on variables
that have higher lags than smaller lags, similar to the approach taken by
the Minnesota Prior. Empirical results show improvement compared to the
baseline BART prior structure and a BVAR. This is joint work with Pedro
Lima, Carlos M. Carvalho and Andrew Herren.

-- 
Hedibert Freitas Lopes, PhD
Professor of Statistics and Econometrics
INSPER - Institute of Education and Research
Rua Quatá, 300 - São Paulo, SP 04546-042 Brazil
Phone: +55 11 4504-2343
www.hedibert.org
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20250319/d3f4a095/attachment.htm>