The knowledge base is a single file encapsulating data
that IBM® Content
Classification requires for accurate
text-based classification. The knowledge base can improve over
time as it learns new categories and adapts to changes in data received
by the system.
Use the Create, Analyze and Learn wizard to build
an effective knowledge base. This feature provides several options:
- Create and analyze knowledge base using active view
- Uses some or all content items to create (train) the knowledge
base, and the remaining part (or all) of the content set to test knowledge
base performance. Choose this option when following a typical knowledge
base creation workflow, as described in Typical workflows. See Creating and analyzing a knowledge base for
instructions.
- Create knowledge base using active view
- Creates a new knowledge base using content items in the active
content set page, without using the Analyze function. See Creating and analyzing a knowledge base for
instructions.
- Analyze knowledge base using active view
- Tests a knowledge base using content items in the active content
set page. You can also choose Analyze knowledge base when
you want to import a knowledge base from a live system, and test it
using other data (that is, items in your content set). See Creating and analyzing a knowledge base for
instructions.
- Learn using active view
- Learning is the capability of the Content Classification to process user feedback
and update the knowledge base accordingly. This option
provides feedback to a knowledge base from content items in the active
content set page. For instructions, see Initiating the learning process.
These options can be performed as a single process
(create and analyze, with learning during analysis) or in several
independent steps (first create, then analyze, then learn), or in
a combination of steps. In addition, you can use the wizard to identify
the predominant language of content items and categories in a knowledge
base (see Language identification).
Another method of building a knowledge base by using keyword data
is also available (see Initializing a knowledge base with keywords).
