ABSTRACT
This chapter provides an introduction to feature selection, the general nomenclature for dimensions reduction methods, and some notable pitfalls. There are a variety of methods to reduce the predictor set. The chapter also provides an overview of the general classes of feature selection techniques. Feature selection methodologies fall into three general classes: intrinsic methods, filter methods, and wrapper methods. The advantage of implicit feature selection methods is that they are relatively fast since the selection process is embedded within the model fitting process; no external feature selection tool is required. Wrappers have the potential advantage of searching a wider variety of predictor subsets than simple filters or models with built-in feature selection. A better way of combining feature selection and resampling is to make feature selection a component of the modeling process. Feature selection should be incorporated the same way as preprocessing and other engineering tasks.
