Hierarchical result views for keyword queries over relational databases

Enabling keyword queries over relational databases (KQDB) benefits a large population of users who have difficulty in understanding the database schema or using SQLs. However, since there are different interpretations for a query, the results of KQDB that mix different possible answers still make it hard for users to consume and extract the information that is interesting to them. To help users with different preferences understand the query results and quickly locate the desired information, this paper proposes generating easyto-navigate hierarchical result views for each interpretation to the query on the fly. We define the structures for organizing these hierarchical views, provide metrics for evaluating them in terms of user navigation efforts, and develop an efficient algorithm for generating the view structures. The effectiveness and efficiency of the proposed approaches are verified through an experimental study. Categories and Subject Descriptors H.2 [Information Systems]: Dat...

• Type: Conference
• Year: 2009
• Where: SIGMOD
• Authors: Shiyuan Wang, Jun'ichi Tatemura, Arsany Sawires, Oliver Po, Divyakant Agrawal, Amr El Abbadi
• Keywords: Database, Hierarchical Views, Retrieval General Terms, SIGMOD 2009, User Navigation Efforts