An overview of penalised regression methods for informing the selection of predictive markers

Abstract

Background: Penalised regression methods are a useful atheoretical approach for identifying key predictive indicators when one’s initial list of indicators is substantial, a process which may aid in informing population health surveillance. The purpose of this study was to examine the predictive performance and feature (i.e., variable) selection capability of common penalised regression methods (LASSO, adaptive LASSO, and elastic-net), compared with traditional logistic regression and forward selection methods. Design: Data were drawn from the Australian Temperament Project, a longitudinal cohort study beginning in 1983. The analytic sample consisted of 1,292 (707 women) participants. A tota..