Object-Oriented XML Keyword Search
No Thumbnail Available
Date
2010-07-22T02:03:58Z
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Keyword search is a user-friendly way to query XML data. Existing LCA-based XML keyword search approaches assign a Dewey ID to every document node and find the relevant LCA nodes of the query keywords based on the Dewey IDs. Despite many improved LCA-based semantics proposed, these approaches are still not as effective as expected.
We observe that object is an important concept for returning meaningful information in database queries. In this paper, we propose an object-oriented approach for XML keyword search.
Our approach only assigns different Dewey IDs to object nodes in the document, and introduces relational tables to organize data values. By considering the semantic relationship among object, property and value during keyword query processing, our approach significantly improves the search efficiency and quality, compared to existing work. Furthermore, after finding any object as a return node, our approach only outputs the useful information about that object, instead of the whole subtree rooted at the object node as in many other approaches. Finally we design an algorithm to rank the possible interpretations of an ambiguous query, and the search results are returned separately based on different query interpretations. The efficiency and effectiveness of our approach are demonstrated with a comprehensive experimental study.