What term describes converting a categorical feature into binary indicators for each category?

• A. Normalization
• B. Dummy Coding
• C. Discretization
• D. Scaling

Answer: (B)

Encoding a categorical feature into binary indicators for each category is called dummy coding. In preprocessing, many models require numeric input, so categorical values are transformed into 0/1 indicators. For a feature with k categories, you create k binary columns, each indicating membership in one category. This approach preserves the distinct identity of each category without implying any ordinal order. For example, a feature like Color with Red, Green, Blue becomes three features: is Red, is Green, is Blue, with a row showing a 1 in the applicable category and 0s elsewhere. If you include an intercept in your model, you might drop one of the columns to avoid perfect multicollinearity, known as the dummy variable trap. Other options like normalization, scaling, or discretization relate to numeric features or converting continuous values into bins, and they don’t turn a categorical feature into binary category indicators.