1 INTRODUCTION Variable selection.

1 INTRODUCTION

Variable selection, or dimension reduction, is fundamental to multivariate statistical standard building. Not only does judicious variable selection improve the model's predictive ability, it also generally provides a better understanding of the underlying universal that generates the data. fit to recent proliferation of large, high-dimensional databases, variable selection has become the focus of intensive research in several areas of that kind as text processing, environmental sciences, and genomics, particularly gene expression array data involving ten or centurys of thousands of variables.

Traditional variable selection approaches, similar as stepwise selection and best subset selection, are built in linear regression