Template-Type: ReDIF-Paper 1.0
Series: Tinbergen Institute Discussion Papers
Creation-Date: 2011-08-19
Number: 11-122/4
Author-Name: Christophe Croux
Author-Email: christophe.croux@econ.kuleuven.be
Author-Workplace-Name: K.U. Leuven, Belgium
Author-Name: Peter Exterkate
Author-Email: exterkate@ese.eur.nl
Author-Workplace-Name: Erasmus University Rotterdam
Title: Sparse and Robust Factor Modelling
Abstract: Factor construction methods are widely used to summarize a large panel of variables by means of a relatively small number of representative factors. We propose a novel factor construction procedure that enjoys the properties of robustness to outliers and of sparsity; that is, having relatively few nonzero factor loadings. Compared to more traditional factor construction methods, we find that this procedure leads to better interpretable factors and to a favorable forecasting performance, both in a Monte Carlo experiment and in two empirical applications to large data sets, one from macroeconomics and one from microeconomics.
Classification-JEL: C38, C51, C53
Keywords: dimension reduction, forecasting, outliers, regularization
File-Url: https://papers.tinbergen.nl/11122.pdf
File-Format: application/pdf
File-Size: 337170 bytes
Handle: RePEc:tin:wpaper:20110122