| Summary: | 碩士 === 國立成功大學 === 資訊管理研究所 === 95 === With the development of Internet, web pages grow rapidly. In
order to search information they need the users often depend on
search engine. A search engine collects web pages in Internet by
information retrieval techniques, and serves as an information
provider to users. However, regarding web pages of the same
concept, the words used by authors and users use may be different.
This is a "word dismatch" problem which prevents users from
retrieving all web pages of the same concept. The solution is
"query expansion"(QE). QE can expand users' queries and let users
gain more complete information.
This research proposes one method for combining ontology with
association rules to perform QE. It can resolve the word dismatch
problem and retrieve more web pages of the same concept and
satisfy users' needs. The method we proposed is based on ontology,
and uses spider to collect web pages as the data set. After the
spider's operation is finished, we will mine the association rules
between words. We provide one QE's recommendation mechanism which
combines words' semantic relationships within ontology with
association rules among words to help user do query.
|
|---|
