Association Thesaurus Construction for Interactive Query Expansion based on Association Rules Mining
碩士 === 國立臺灣科技大學 === 電子工程系 === 89 === World Wide Web has become a major medium for information publication and communication. People either browse or use the search service when they want to find specific information on the web. The popular way to find desired information is using major search engi...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2001
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/40842963601965375088 |
| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 89 === World Wide Web has become a major medium for information publication and communication. People either browse or use the search service when they want to find specific information on the web. The popular way to find desired information is using major search engines. Vocabulary problem are raised on information retrieval issues. Our approach to these problems is to provide users with query expansion based on association thesaurus. To construct association thesaurus, association rule mining is employed to generate asymmetric relationship between terms by relevance feedback. It is basis of assuming that rated web pages on every search result are correlated even though some rated web pages are not desired for the user. Hence, rated web pages are transferred to user-interested terms based on terms co-occurrence by ATF-DF measure we propose. Then these list of user-interested terms are mined to produce association rules. Interactive query expansion uses structural query language to respond candidate expansion terms for user selecting from association thesaurus. In addition, novel method whose idea comes from cooperative process of self-organizing map expands candidate expansion terms when rare candidate expansion terms. Experimental results show that query logs are used to transfer to knowledge is feasible. The quality and completeness of association thesaurus depend on the query logs collection greatly. We apply interactive query expansion with association thesaurus to the Query Agent on the Coursebot which is web-based intelligent recommendation system for e-learning. The Query Agent has improvement of retrieval effectiveness on the Coursebot. It can increase precision rate of search without poor recall rate.
|
|---|