Fact Name

Fact Name represents the name of the facts that we will extract. The Fact Names can be arranged hierarchically into taxonomies. This can help us represent, at an abstract level, the hierarchical or containment structure of the actual facts being extracted. Typically, a leaf Fact Name will represent the lowest level of fact or data, whereas a parent Fact Name will represent groupings of those low level facts, and, indeed, groupings of groupings, and so on.

• We represent the names and abstract structure of the facts and fact groups in the “FACT NAMES” Concept Scheme.

• If you look at the screenshot, we can see we have a concept representing the Fact Name “Provenance,” and that it has two child Fact Names: “Chef / Cook” and “Cookbook.”

• This is how we would like our facts about Provenance to be returned to us, with two children, one for the Chef / Cook fact, and the other for the Cookbook fact.

• The Fact Names will appear as the Class names returned by Classification Server when tagging documents

• Of course, ultimately, we have to build fact extractors that return us that structure – this is just the “fact design” or “fact schema” we would like!