A wavelet-based multivariate multiscale approach for forecasting

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

In our increasingly data-rich environment, factor models have become the workhorse approach for modelling and forecasting purposes. However, factors are not observable and have to be estimated. In particular, the space spanned by the unknown factors is typically estimated via principal components. This paper proposes a novel procedure for estimating the factor space, resorting to a wavelet-based multiscale principal component analysis. A Monte Carlo simulation study is used to demonstrate that such an approach may improve both the estimation and the forecasting performances of factor models. The empirical application then illustrates its usefulness for forecasting GDP growth and inflation in the United States.

Original languageEnglish
Pages (from-to)581-590
Number of pages10
JournalInternational Journal of Forecasting
Volume33
Issue number3
DOIs
Publication statusPublished - 1 Jul 2017

Keywords

  • Factor models
  • Forecasting
  • Multiscale principal components
  • Wavelets

Fingerprint Dive into the research topics of 'A wavelet-based multivariate multiscale approach for forecasting'. Together they form a unique fingerprint.

Cite this