Skip to content

Week #5 – Features vs. Variables

The machine learning community calls them features (also attributes or inputs).  There is a subtle difference in meaning.  In predictive modeling, depending on the nature of the data, considerable work may be required to transform and winnow down data and text to a usable set of predictors.  This process is termed “feature engineering.”  The comparable term in statistics – “variable selection” – does not imply such a wide ranging process.