In this paper, we compare two different variable selection approaches for linear regression models: Autometrics (automatic general-to-specific selection) and LASSO (?1-norm regularization). In a simulation study, we show the performance of the methods considering the predictive power (forecast...