Autor: MAURICIO FRANCA LILA
Orientadores: Fernando Luiz Cyrino Oliveira & Erick Meira de Oliveira
Data e Hora: 28/04/2023, 10:30h
Link/ Sala: https://puc-rio.zoom.us/j/92183949231?pwd=NllydW5pTDlLNFltN0pIN2FQa1JCdz09
Banca Examinadora: Fernando Luiz Cyrino Oliveira – orientador – PUC-Rio; Erick Meira de Oliveira – co-orientador -FINEP; Helio Côrtes Vieira Lopes – PUC-Rio; Reinaldo Castro Souza – PUC-Rio; Lilian Manoel de Menezes Willenbockel – UL; Lupercio França Bessegato – UFJF; Paulo Jorge Canas Rodrigues – UFBA.
Resumo:
This study presents a set of methodological proposals related to the forecast reconciliation
in the context of Hierarchical Time Series. The main objective is to present original solutions to the theme, seeking to obtain more accurate forecasts than those obtained by independent models for the different levels of the hierarchy. The studies were conducted in real data, showing the potentiality of application of the methods developed in different scenarios, where the time series are structures in a hierarchical fashion. This thesis is composed of a set of essays that explore forecast reconciliation from theperspective of a regression model, which gives foundations to optimal reconciliation.The first contribution addresses the problem of forecast reconciliationfrom the perspective of robust estimators. The proposal presents an original contribution applied to data from labor force surveys in Brazil, presenting a set of solutions that can drive efficient public policies. In this case, the reconciledforecasts obtained through robust estimators provided consistent gains in terms of accuracy when compared to methods that represent the state-of-the-art onforecast reconciliation in hierarchical time series. The second contribution deals with the problem of optimal reconciliation applied to energy consumption seriesin Brazil. We presented an alternative proposal, less sensitive to outlyingforecasts at the reconciliation stage. The results obtained in this second study show considerable improvements in standard evaluation metrics with regard to the new forecasts. A third proposal seeks to offer robust covariance structuresof forecasting errors, which expands the set of strategies presented in the literature.The main contribution is to incorporate robust covariance estimatesinto the MinT (Minimum Trace) reconciliation approach, which minimizes reconciliation errors, offering an estimator with minimum variance.