ML//feature selection

Feature selection is the step of model building that cuts a candidate feature set down to the few features that carry signal, done before an evaluation is trusted.

Feature selection is the step of model building that cuts a candidate feature set down to the few features that carry signal, done before an evaluation is trusted.

In rare-event problems it is debt control: every extra feature is a loan against evidence the data probably does not hold. It fights the curse of dimensionality.

Redundant features (two ways of saying the same thing) get dropped by correlation and by meaning: keep one reading, drop synonyms.

Fewer features means a model small enough that a human can still argue with it, and less room for overfitting.