You can organize your knowledge base by using a hierarchical structure. For example, you might want to organize your knowledge base by languages, and then organize sub-branches by departments in your company.
You can build a hierarchical knowledge base as you wish, and use rules where necessary. IBM® Content Classification handles statistical issues internally. The only restriction is in the placement of rule nodes.
The most common type of hierarchical knowledge base distinguishes between the levels based on deterministic criteria. In this type of knowledge base, some of the higher-level nodes have rules. These rules enable the system to direct items to a specific set of dependent categories, based on values that are matched in selected fields or other criteria such as language, empty body, and so on. For example, if you need to separate sets of categories according to the source of incoming items, you can create a field that discerns the source (for example, a particular web page, or a secured vs. unsecured web channel). Then you can use the value in a rule associated with the parent nodes of each subset of categories (see instructions for adding nodes,).
To take advantage of the ability of the Content Classification to analyze items in multiple languages, you must build a multilingual knowledge base. This knowledge base can be hierarchical, as explained in Using rules to differentiate between category sets.