Affiliations: [a] College of Software, Northeastern University, China
| [b] College of Computer Science and Engineering, Northeastern University, China
Correspondence:
[*]
Corresponding author. Jingwei Cheng, College of Computer Science and Engineering, Northeastern University, China. Tel./Fax: +86 024 83681582; E-mail: chengjingwei@mail.neu.edu.cn.
Abstract: In recent years there have been a great deal of interests in effectively extracting information from RDF repositories. However, users often experience empty answer problems (EAP) when they are not familiar with the structure or content of a RDF dataset and issue a restrictive query to the system. A solution is to allow users to incorporate fuzzy conditions into the query to make it more flexible. Although a fuzzy query is more or less satisfactory, there are still occasions when no or too few entities in a RDF dataset satisfy the query criteria with a degree more than zero, which is called a fuzzy-EAP problem. In this paper, we propose a query relaxation approach based on relative proximity relations to tackle this problem. We present a framework of fuzzy query relaxation over RDF datasets, and present query relaxation algorithms for atomic f-SPARQL queries and complex f-SPARQL queries. Moreover, this paper reports on a prototype implementation on an extension of LUBM dataset to verify the feasibility and effectiveness of the proposed method. The experimental results show a relatively high recall as more relevant results are returned after relaxation.
