The workshop includes presentations by:

• Florian Huber, University of Salzburg: 

  Title: "A Nonparametric Approach to Augmenting a Bayesian VAR with Nonlinear Factors" This paper introduces a Bayesian VAR with nonlinear factors modeled via regression trees. It reduces mis-specification risk, enables efficient MCMC estimation, and adapts structural shock identification to nonlinear settings, showing strong forecasting and analytical performance.

• Christiane Baumsteiner, University of Notre Dame, South Bend, Indiana:
  Title: "Oil, Inflation Expectations, and Household Characteristics: A Nonlinear Heterogeneous Agent VAR Approach"

  This paper develops a scalable micro-macro framework combining linear time series and nonlinear panel models to study how oil supply shocks affect inflation expectations across Euro area households, revealing strong asymmetries and heterogeneity by country, gender, income, and age.

• Davide Ferrari, unibz
  Title: "Robust Anomaly Detection in High-Dimensional Heterogeneous Data" This talk introduces a framework for outlier detection in settings with a very large number of mixed-type features (continuous, categorical, ordinal, and sparse). We develop a mixed-data influence measure that provides control of false discoveries under heavy tails and complex dependence structures, supported by scalable algorithms and rigorous theoretical guarantees. Applications include financial risk monitoring, fraud detection in banking transactions, and business process analytics, where the method uncovers structured anomalies that remain hidden to standard low-dimensional approaches