User’s Guide : Basic Single Equation Analysis : Regression Variable Selection
  
Regression Variable Selection
This chapter describes tools that may be used to automatically determine the variables used as regressors in a least squares regression. Variable selection, or feature selection, is an important component of modern data analysis.
The explosion in available data over recent decades, coupled with increases in computing power, has led to growing popularity of methods that allow the data themselves to suggest the most appropriate combination of regressors to use in estimation. These techniques allow the researcher to provide a set of candidate variables for the model, rather than specifying a specific model.
The chapter begins with some general discussion concerning variable selection in EViews, then describes three variable selection techniques available, and concludes with some examples:
• “Variable Selection Methods” in EViews provides an introduction to variable selection in EViews.
Each of the supported methods is described in some detail:
• “Uni-Directional”.
• “Stepwise”.
• “Swapwise”.
• “Combinatorial”.
• “Auto-Search / GETS” outlines an improvement of stepwise regression techniques that adds tests of model validity to the selection process.
• “Lasso” describes the use of Lasso estimation (“Chapter 36. Elastic Net and Lasso”) as a tool for least squares variable selection.