Document classifier (TAXON) [7] is used to generate
categories of documents retrieved from the initial user
natural language query. TAXON represents the documents
as a reduced vector, identifies the relevance of the vector
about documents and query subjects, and then classifies the
documents. By using a thesaurus tools, it also enables
document classification based on concepts by acquiring
closely related meanings with the subjects of document.
Table 1 shows the performance (classification power) of the
TAXON used in experiment.