Features roles and types¶
A feature’s role determines how it’s used during machine learning.
Reject means that the feature is not used
Use for display only means that the feature is not used to build a model, but is used to label model output. This role is currently only used by cluster models.
A feature’s variable type determines the feature handling options during machine learning.
Categorical variables take one of an enumerated list values. The goal of categorical feature handling is to encode the values of a categorical variable so that they can be treated as numeric.
Numerical variables take values that can be added, subtracted, multiplied, and so on. There are times when it may be useful to treat a numerical variable with a limited number of values as categorical.
Text variables are arbitrary blocks of text. If a text variable takes a limited number of values, it may be useful to treat it as categorical.
Vector variables are arrays of numerical values, of the same length.
Image variables are available for Deep learning. See Using image features for more information.