Knowledge Maxxing← Back to topic library
ML MaxxingIntermediate75 sec
One-hot encoding
One-hot encoding converts a categorical feature into several binary columns, one per possible category, each marking whether that category applies to a given example.
Prerequisite: Feature scaling
What matters
Most models expect numeric input and would otherwise misread category labels as having an order or magnitude, so one-hot encoding avoids implying a false ranking between categories.
Example
A 'color' feature with values red, blue and green becomes three separate 0/1 columns: is_red, is_blue, and is_green.
Common misconception
One-hot encoding does not work well for categorical features with thousands of possible values; it can produce an impractically large number of columns.
Recall question
Why can't most models directly use a text category label as a plain number?
Keywords and related concepts
one-hot encodingcategorical variablepreprocessingfeature engineering