Feature Selection
noun
say it: “FEE-cher sih-LEK-shun” /ˈfitʃər səˈlɛkʃən/
Choosing which input columns a model should use and leaving the rest out.
What is feature selection?
Feature selection is choosing which features a model gets, such as keeping the ten most related to the label and dropping the rest. When the choosing looks at the labels, it belongs inside the pipeline, on the training rows only; done on every row before splitting, it peeks at the test answers and leaks.
In a sentence
- “Feature selection kept 12 of the 300 columns.”
- “Doing feature selection before the split let the test labels choose the features.”
Level: Journeyman · More AI Basics terms at this level
Updated