Automatic Extraction of a Domain-Oriented Thesaurus from Manually Classified Documents