This research used the upper n (30) documents to
generate classification groups. The procedure of generating
classification groups and term cluster for the upper 30
documents is as follow: (1) acquire classification information
through document classifier for the top 30 retrieved
documents, (2) group the documents that contain the same
classification information by checking the acquired
classification information from (1), (3) generate term cluster
by extracting terms that represent each classified group, and
(4) have users to select appropriate term cluster.