The Community for Technology Leaders
2014 International Conference on Big Data and Smart Computing (BIGCOMP) (2014)
Bangkok, Thailand
Jan. 15, 2014 to Jan. 17, 2014
ISBN: 978-1-4799-3919-0
pp: 1-6
Tok Wang Ling , School of Computing, National University of Singapore, Singapore
Thuy Ngoc Le , School of Computing, National University of Singapore, Singapore
Zhong Zeng , School of Computing, National University of Singapore, Singapore
ABSTRACT
Keyword search has been the major form of retrieval method in information retrieval system, and has become an important way for novice to explore data-centric XML and relational databases (RDB). Recent years have witnessed many approaches proposed for keyword search over XML and RDB. However, those approaches cannot intelligently exploit hidden semantics in XML or RDB, and thus encounter serious problems in processing keyword queries. In this paper, we point out mismatches between query answers returned by existing approaches and the common expectations in keyword search over XML and RDB. We analyze these mismatches and discover that the main reasons are due to the unawareness of semantics of object, relationship and attribute in databases. To capture these semantics, we construct Object Relationship (OR) data graph for XML and Object Relationship Mixed (ORM) data graph for RDB, and propose an intelligent keyword search based on OR and ORM data graph model to retrieve more informative answers. Finally, to further facilitate the usability of keyword search, we also show our ongoing work to enhance the expressive power of keyword queries. Particularly, we 1) enable users to explicitly indicate their search intentions by relation, attribute and tag names in keyword queries; 2) handle recursive relationships and identifier-dependency relationships (IDD) in databases; and 3) incorporate aggregate function into keyword queries so that users can explore databases with aggregate queries.
INDEX TERMS
CITATION
Tok Wang Ling, Thuy Ngoc Le, Zhong Zeng, "Towards an intelligent keyword search over XML and relational databases", 2014 International Conference on Big Data and Smart Computing (BIGCOMP), vol. 00, no. , pp. 1-6, 2014, doi:10.1109/BIGCOMP.2014.6741406
95 ms
(Ver 3.3 (11022016))