Inference in Bayesian additive vector autoregressive tree models

DOI10.1214/21-AOAS1488zbMATH Open1498.62309arXiv2006.16333OpenAlexW3040075297MaRDI QIDQ2135338FDOQ2135338

Authors: Florian Huber, Luca Rossini 

Publication date: 6 May 2022

Published in: The Annals of Applied Statistics (Search for Journal in Brave)

Abstract: Vector autoregressive (VAR) models assume linearity between the endogenous variables and their lags. This assumption might be overly restrictive and could have a deleterious impact on forecasting accuracy. As a solution, we propose combining VAR with Bayesian additive regression tree (BART) models. The resulting Bayesian additive vector autoregressive tree (BAVART) model is capable of capturing arbitrary non-linear relations between the endogenous variables and the covariates without much input from the researcher. Since controlling for heteroscedasticity is key for producing precise density forecasts, our model allows for stochastic volatility in the errors. We apply our model to two datasets. The first application shows that the BAVART model yields highly competitive forecasts of the US term structure of interest rates. In a second application, we estimate our model using a moderately sized Eurozone dataset to investigate the dynamic effects of uncertainty on the economy.

Full work available at URL: https://arxiv.org/abs/2006.16333

Recommendations

• Bayesian nonparametric vector autoregressive models
• Large Bayesian vector autoregressions with stochastic volatility and non-conjugate priors
• Nonparametric vector autoregressions: specification, estimation, and inference
• Bayesian Vector Autoregressions with Stochastic Volatility
• scientific article; zbMATH DE number 1098833

zbMATH Keywords

nonparametric regressiondecision treesvector autoregressionsBayesian additive regression treesBAVART

Mathematics Subject Classification ID

Bayesian inference (62F15) Nonparametric regression and quantile regression (62G08) Time series, auto-correlation, regression, etc. in statistics (GARCH) (62M10) Applications of statistics to economics (62P20) Economic time series analysis (91B84)

Cites Work

• A decision-theoretic generalization of on-line learning and an application to boosting
• Prediction with missing data via Bayesian additive regression trees
• BART: Bayesian additive regression trees
• Greedy function approximation: A gradient boosting machine.
• Bayesian additive regression trees using Bayesian model averaging
• Ancillarity-sufficiency interweaving strategy (ASIS) for boosting MCMC estimation of stochastic volatility models
• Forecasting the term structure of government bond yields
• Random forests
• Bayesian treed Gaussian process models with an application to computer modeling
• Forecasting and conditional projection using realistic prior distributions
• Bayesian regression tree models for causal inference: regularization, confounding, and heterogeneous effects (with discussion)
• Bayesian treed models
• Bayesian nonparametric sparse VAR models
• Bayesian Nonparametrics
• Impulse response analysis in nonlinear multivariate models
• Title not available (Why is that?)
• The Impact of Uncertainty Shocks
• Beta-product dependent Pitman-Yor processes for Bayesian inference
• The nonlinear effects of uncertainty shocks
• Real-time density forecasts from Bayesian vector autoregressions with stochastic volatility
• Bayesian nonparametric vector autoregressive models
• Bayesian regression trees for high-dimensional prediction and variable selection
• Bayesian compressed vector autoregressions
• Large Bayesian vector autoregressions with stochastic volatility and non-conjugate priors
• Automated versus do-it-yourself methods for causal inference: lessons learned from a data analysis competition
• Heteroscedastic BART via Multiplicative Regression Trees
• Bayesian inference on structural impulse response functions
• Corrigendum to ``Large Bayesian vector autoregressions with stochastic volatility and non-conjugate priors
• Nonlinear impacts of international business cycles on the U.K. -- a Bayesian smooth transition VAR approach

Cited In (6)

• Bayesian mixed-frequency quantile vector autoregression: eliciting tail risks of monthly US GDP
• TAIL FORECASTING WITH MULTIVARIATE BAYESIAN ADDITIVE REGRESSION TREES
• Bayesian additive regression trees with model trees
• Bayesian nonparametric vector autoregressive models
• Nonparametric vector autoregressions: specification, estimation, and inference
• Real-time density forecasts from Bayesian vector autoregressions with stochastic volatility

Uses Software

• AdaBoost.MH
• BART-BMA
• BartPy
• XBART

This page was built for publication: Inference in Bayesian additive vector autoregressive tree models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2135338)