Combining Factor Models and Variable Selection in High-Dimensional Regression

Abstract

This presentation provides a summary of some of the results derived in Kneip and Sarda (2011). The basic motivation of the study is to combine the points of view of model selection and functional regression by using a factor approach. For highly correlated regressors the traditional assumption of a sparse vector of parameters is restrictive. We therefore propose to include principal components as additional explanatory variables in an augmented regression model.