Concept Based Query Enhancement in the ARCH Search Agent

The quality of the search experience has been an enduring problem for the World Wide Web. One of the well-known difficulties is the tendency of users to use short, under-specified and ambiguous queries, which tend to retrieve large amounts of irrelevant material. Traditional query expansion and relevance feedback approaches have been used to address this problem, but without as much success as these techniques have garnered in more restricted areas such as traditional IR collections. This paper presents ARCH, an interactive query formulation aid that is based on conceptual categories. The user’s query is reformulated to include categories that the user recognizes as important and exclude those that are not important. Unlike query expansion techniques, which might add lists of synonyms to increase recall, ARCH uses the domain knowledge inherent in Web-based classification hierarchies such as Yahoo to add just those terms likely to improve the match with the user’s intent. The goal...

• Type: Conference
• Year: 2003
• Where: IC
• Authors: Ahu Sieg, Bamshad Mobasher, Steven L. Lytinen, Robin D. Burke
• Keywords: IC 2003, IC 2007, Query Expansion, Query Formulation Aid, Traditional Query Expansion