statement, and its normalized logic expression can be
gotten. Thirdly, the initial concept set is obtained through
keywords extracting. Fourthly, relevant precise concepts
and rough concepts which can satisfy the threshold are
determined by rough ontology based semantic similarity
calculation. Fifthly, an extended concept set is fixed by
semantic extending with the initial concept set. Sixthly,
the extended concept set is used to match the documents
annotated in semantic indexed document database. Lastly,
the matched documents are ranked into query result list,
and then obtained document set as the final result is
returned to the user.