You can make the associations or the sequences that are
found among items more meaningful if you group the items in categories.
You can group these categories again into subcategories. The result
is a hierarchy of categories with the items on the lowest level. This
is called a taxonomy.
You can use taxonomies with the Associations mining function or
with the Sequence Rules mining function.
The following example illustrates a taxonomy with three hierarchy
levels and a balanced hierarchy:
Figure 1. A taxonomy with a balanced hierarchy Level 0 Level 1 Level 2
Orange Juice |
>> Beverages |
Beer | |
|
>> Food
|
Tomato | |
| |
Cabbage >> Vegetables |
|
Carrot |
If you apply a taxonomy during an Associations mining run or a
Sequence Rules mining run, the groups are expanded for the calculation
of the rules. If an item of a group is a member of a category, this
category is added to the group. If the added category is a member
of another category, this other category is also added to the group,
and so on. During the calculation of the rules, the categories in
the expanded groups are handled in the same way as the items. Therefore,
the rules refer not only to the items but also to the categories.
For example, the rule:
might be accompanied by the following rules:
- [Beverages] => [...]
- [Food] => [...]
This means that a taxonomy enables different items or categories
to be considered as equivalent under a new category name. An item
or a category can be a member of one category or more categories,
or of none.