A nominal variable, also known as a categorical variable, expresses the membership of a data item in one or more discrete groups.
It is used when there is a need to classify data into distinct categories without any intrinsic order. Nominal variables are commonly applied in scenarios such as demographic data, survey responses, and classification tasks. The technique works by assigning labels to each category, allowing models to understand the distinct groups within the data.
In constrast to ordinal categories, nominal variables cannot be order in a reasonable way.
For example, in a survey, gender is represented as nominal variable.
Similarly, countries like Germany
, France
, and UK
should be represented as nominal variables.
Nominal variables are often contrasted with quantitative variables, which are represented by numerical values. In some cases, classifications are internally represented by probabilities, such as prediction probabilities or confidence scores, which indicate the likelihood of each category.
Nominal variables are important because they provide a way to represent categorical data, enabling models to process and analyze data where the categories are distinct and unordered. If data points can have more than one assigned category at the same time, this is referred to as a multi-label or multi-output setting.
- Alias
- Related terms
- Classification Vector of Categorical Variables Ordinal Variable Prediction Probability