Combination of DM and OBDA for Ontology Generation from Relational Database

Wenge Rong,Rifan Li,Hao Sheng,Yuanxin Ouyang,Zhang Xiong
DOI: https://doi.org/10.1109/DASC.2013.94
2013-01-01
Abstract:Nowadays a large amount of data on the web are stored in relational database (RDB). How to effectively map these data to ontology has become one of the most important challenges in semantic web. Several efforts have been conducted in this area and direct mapping (DM) is a notable one as the W3C candidate recommendation. Though DM has shown its effectiveness, it still faces a lot of challenges. Its performance in term of storage space and time complexity will become worse when RDBS is becoming big and updating frequently. Furthermore, it cannot satisfy monotonicity and semantics preservation at the same time. In this paper, inspired by the idea of ontology-based data access (OBDA), a mechanism is proposed where database is accessed via description logic (DL) ontology and a real TBox and mapping file will be generated accordingly. This method can achieve similar function compared to standard ontology like OWL. The experimental study has shown that it can provide better performance with regard to the storage space and ontology generation time. It has also shown its capability in supporting rapid query process. Different from the other methods which generate virtue RDF data, the proposed approach can also support adding rules, which will enhance the semantic expression ability of the knowledge base.
What problem does this paper attempt to address?