DTA

Digital Theses Archive

 

Tesi etd-10272022-181837

Type of thesis
Corso Ordinario Secondo Livello
Author
CIANCETTA, ALESSANDRO
URN
etd-10272022-181837
Title
Heterogeneous Responses to Structural Shocks in Deep Dynamic Factor Models
Structure
Cl. Sc. Sociali - Scienze Economiche
Course
SCIENZE ECONOMICHE E MANAGERIALI - SCIENZE ECONOMICHE E MANAGERIALI
Committee
Relatore Prof. MONETA, ALESSIO
Presidente Prof. PICCALUGA, ANDREA MARIO CUORE
Tutor Prof. FAGIOLO, GIORGIO
Membro Prof. NUVOLARI, ALESSANDRO
Membro Dott.ssa CANTARELLI, PAOLA
Membro Prof. MINA, ANDREA
Membro Prof. ROVENTINI, ANDREA
Keywords
  • Nessuna parola chiave trovata
Exam session start date
02/12/2022;
Availability
parziale
Abstract
Deep learning is a powerful tool for obtaining complex representations of data. In this work, we consider the model proposed by Andreini, Izzo and Ricco (2020), which applies deep autoencoders to the estimation of non-linear factor models. We extend their theoretical framework to study the effects of structural shocks on the system. We focus on the possibility of heterogeneous responses depending on the initial state of the economy at the time when the shock occurs, and we propose a new estimation method for the Impulse Response Functions (IRFs) in this context. We test our method on simulated data and then present an application using the FRED-MD dataset.
Files