A1308
Title: Model selection in large dimensional linear regression using sequential multiple testing
Authors: Vasilis Sarafidis - Brunel University London (United Kingdom) [presenting]
George Kapetanios - Kings College London (United Kingdom)
Alexia Ventouri - Kings College London (United Kingdom)
Abstract: High-dimensional regression specification and analysis is a complex and active area of research in statistics and econometrics. A large number of approaches have been proposed, but determining their relative merits remains a challenging task. The aim is to propose a new hybrid approach. It combines elements from two existing methods. The first is the greedy methodology developed in a prior study, where a powerful multiple testing step introduces parsimony by ensuring that completely irrelevant variables are not selected with high probability. The second is stage-wise regression, where relevant variables are selected in steps, rather than jointly as in the prior study. However, in that literature, the stopping rules, which typically rely on model information criteria, are not sufficiently parsimonious. Some theoretical properties of the new method are derived and shown, through simulations, to perform well. An illustration, using corporate emissions data, provides an empirical perspective.
