Deriving a Conceptual Search Ontolection from a Domain Collection
As discussed in Overview, Watson™ Explorer Engine provides two types of conceptual search: thesaurus-driven conceptual search, which uses ontolections that you must define and crawl, and data-driven conceptual search, which automatically suggests alternate words based on a domain collection and either a wildcard dictionary created from domain collection or a domain list containing industry specific words. This tutorial has started with thesaurus-driven conceptual search and we now turn to the automatic generation of suggested terms. In order for data-driven conceptual search to generate effective suggestions based on the best possible analysis of your search results, the domain collections used should be large enough to return a relatively large number of results. This is not the case with the collection used in this tutorial. However, we can use this small collection to explain how to configure data-driven conceptual search for your search collections.
This following sections explain how to derive a conceptual search ontolection from a domain collection.
To proceed with this tutorial, click First Steps in Deriving a Conceptual Search Ontolection.