Subspace methods

Research output: Chapter in Book/Report/Conference proceedingChapter (Book)Otherpeer-review

Abstract

With increasingly many variables available to macroeconomic forecasters, dimension reduction methods are essential to obtain accurate forecasts. Subspace methods are a new class of dimension reduction methods that have been found to yield precise forecasts when applied to macroeconomic and financial data. In this chapter, we review three subspace methods: subset regression, random projection regression, and compressed regression. We provide currently available theoretical results, and indicate a number of open avenues. The methods are illustrated in various settings relevant to macroeconomic forecasters.

Original languageEnglish
Title of host publicationMacroeconomic Forecasting in the Era of Big Data
Subtitle of host publicationTheory and Practice
EditorsPeter Fuleky
Place of PublicationCham Switzerland
PublisherSpringer
Pages267-291
Number of pages25
Edition1st
ISBN (Electronic)9783030311506
ISBN (Print)9783030311490
DOIs
Publication statusPublished - 2020

Publication series

NameAdvanced Studies in Theoretical and Applied Econometrics
Volume52
ISSN (Print)1570-5811
ISSN (Electronic)2214-7977

Cite this

Boot, T., & Nibbering, D. (2020). Subspace methods. In P. Fuleky (Ed.), Macroeconomic Forecasting in the Era of Big Data: Theory and Practice (1st ed., pp. 267-291). (Advanced Studies in Theoretical and Applied Econometrics; Vol. 52). Springer. https://doi.org/10.1007/978-3-030-31150-6_9