How Categorizer Works

• A user (administrator) creates a catalogue structure and defines the list of categories.
• Etalon documents are selected for each category / subcategory on the base of which the Categorizer forms a fingerprint. Each category corresponds to the fingerprint.
  Fingerprint - is a set of main concepts (or terms) and relations between them, extracted from the etalon categories.
• Documents for categorization are parsed by Linguistic Processor which extracts syntactic and semantic relations from them.
• Documents presented in the form of semantic tree are compared with the fingerprints of each category. As a result of comparison, Intellexer Categorizer estimates every document proximity degree to each category.
• It is determined for each document:
  • High relevant categories
  • Low relevant categories
  • Irrelevant categories

In order to learn the principles of how Intellexer Categorizer functions, please visit Online Demo.

To get more information about Intellexer Categorizer read also:

• Benefits and Features of Intellexer Categorizer
• The description of linguistic processor operation principles