Within the context of multilevel categorization, when you create a categorization schema, you can choose whether you want categories arranged hierarchically or by attributes.
Hierarchical categorization
You can choose the hierarchical categorization in the category modeler.
This is a strict hierarchical arrangement of categories in which each category completely describes a subject. Higher-level categories merely represent the context of each subject.
You cannot use any category duplicates in the hierarchical categorization.
Example of a hierarchical categorization:
Groceries
Packaging
Stable
Unstable
Taste
Delicious
Unpleasant
Attribute categorization
You can choose attribute categorization in the category modeler.
Attribute categorization is not strictly hierarchical and allows you to use category duplicates to map value combinations.
With attribute categorization, each category is a characteristic of a subject, without which a subject would not be fully described. A subject is described by using a combination of several categories. Such a combination corresponds to a path in the category hierarchy. Shared properties of different subjects are mapped in the form of category duplicates.
Example of an attribute categorization:
Groceries
Packaging
Good
Bad
Taste
Good
Bad
In this example,
Good
and
Bad
are category duplicates, or semantically identical categories. In general, the following subjects are expressed here as combinations of characteristics:
{Groceries, packaging, good}
{Groceries, packaging, bad}
{Groceries, taste, good}
{Groceries, taste, bad}