Non-ordered nominal attributes can be recoded using which coding scheme?

• A. One-Hot Encoding
• B. Binary Coding
• C. Dummy coding
• D. Label Encoding

Answer: (C)

When a variable is nominal with no natural order, you want to represent each category by a simple on/off signal rather than a numeric value that could imply ranking. Dummy coding does precisely this by turning the categories into binary indicators. For a variable with k categories, you create k–1 binary features (one category serves as the baseline). Each feature marks whether an observation belongs to that category, with zeros otherwise. This keeps the data purely categorical in effect—no order is suggested—and it works well with regression and many machine learning models because it avoids implying any magnitude or ranking among categories.

One-hot encoding is essentially the same idea in practice, just described from a slightly different angle, and both avoid ordinal misinterpretation. Label encoding, by assigning integers to categories, can mislead models into thinking there is an meaningful order. Binary coding uses binary representations of category labels and can also encode artificial ordinality. So for non-ordered nominal attributes, dummy coding is the appropriate approach.