Feature
An individual measurable input variable used by a model to make predictions.
- Columns of the input matrix X.
- Can be numeric, categorical, text, pixels, or engineered.
- Feature design often matters more than algorithm choice.
When to use: Every supervised ML problem; designing and selecting features is most of classical-ML practice.
Example: House price model features: square_footage, num_bedrooms, zip_code, year_built, and engineered features like price_per_sqft_in_zip.