Advanced search
Start date
Betweenand

Bayesian semiparametric analysis of autoregressive models

Grant number: 17/10096-6
Support type:Scholarships in Brazil - Doctorate
Effective date (Start): August 01, 2017
Effective date (End): February 29, 2020
Field of knowledge:Physical Sciences and Mathematics - Probability and Statistics
Cooperation agreement: Coordination of Improvement of Higher Education Personnel (CAPES)
Principal Investigator:Hedibert Freitas Lopes
Grantee:Helton Graziadei de Carvalho
Home Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
Associated scholarship(s):17/22914-5 - Time-clustering and forecasting performance in semi-parametric INAR(1) models, BE.EP.DR

Abstract

Autoregressive models for time series have been extensively studied during the past several decades, generally motivated by the AR Model or the INAR Model. For instance, several papers generalize the innovation distribution or thinning operator for the INAR(1) model. However, we propose in this project a way to cluster the innovation rates introducing a Dirichlet Process Prior on the time-varying innovations. Furthermore, we show a proposition which is useful to obtain the predictive distribution for the generalized model. The proposed methods will be applied to real datasets. (AU)