Cross: an OWL wrapper for teasoning on relational databases

Cross: an OWL wrapper for teasoning on relational databases
Cross: an OWL wrapper for teasoning on relational databases
Bibliographical Metadata
Year:	2007
Authors:	Pierre-Antoine Champin, Geert-Jan Houben, Philippe Thiran
Venue	ER
Content Metadata
Problem:	Transforming Relational Databases into Semantic Web
Approach:	No data available now.
Implementation:	Cross
Evaluation:	No data available now.

Abstract

One of the challenges of the Semantic Web is to integrate the huge amount of information already available on the standard Web, usually stored in relational databases. In this paper, we propose a formalization of a logic model of relational databases, and a transformation of that model into an OWL, a Semantic Web language. This transformation is implemented in Cross, as an open-source prototype. We prove a relation between the notion of legal database state and the consistency of the corresponding OWL knowledge base. We then show how that transformation can prove useful to enhance databases, and integrate them in the Semantic Web.

Conclusion

In this paper, we have proposed the ODBC model, a formalization of relational databases focusing on their logic model. We have then presented a transformation of that model into OWL, a DL-based language designed for the Semantic Web. This transformation is implemented by the Cross open-source prototype, which effectively introduces the interesting notion of semantic values. We proved that the knowledge-based produced by this transformation is consistent if and only if the source database state is weakly legal (i.e. legal but regarding foreign key constraints). Taking advantage of that result, we have shown how that transformation can prove useful for the purpose of analysing legacy RDBs, enhancing existing RDBs with additional constraints, and integrating them in the SW.

Future work

A first direction for further work would be to try and strengthen the theorem, to have an equivalence of OWL consistency with full legality, i.e. taking into account foreign keys. This could actually be done by using an expressive feature of OWL (the oneOf constructor, not mentioned in this paper), but would possibly make the reasoning intractable. Another solution would be to propose, in a similar way to finite model reasoning, an algorithm of closed world reasoning which would not be allowed to create individuals. We also want to get more experimental results for the Cross implementation. Preliminary results 7 are encouraging: the transformation of the schema of real database (127 tables, 869 columns, 132 unicity constraints, no foreign key) took around 1.5s; the resulting ontology was loaded in Pellet in about 9s, while reasoning took about 3s. Those results seem reasonable for a quite big schema. We now plan to experiment on the use cases presented in Section 6.3 with that database and a sample of other real databases.

Approach

Positive Aspects: No data available now.

Negative Aspects: No data available now.

Limitations: No data available now.

Challenges: No data available now.

Proposes Algorithm: No data available now.

Methodology: No data available now.

Requirements: No data available now.

Limitations: No data available now.

Implementations

Download-page: http://liris.cnrs.fr/∼pchampin/dev/cross

Access API: No data available now.

Information Representation: No data available now.

Data Catalogue: {{{Catalogue}}}

Runs on OS: No data available now.

Vendor: No data available now.

Uses Framework: No data available now.

Has Documentation URL: No data available now.

Programming Language: No data available now.

Version: No data available now.

Platform: No data available now.

Toolbox: No data available now.

GUI: No

Research Problem

Subproblem of: No data available now.

RelatedProblem: No data available now.

Motivation: No data available now.

Evaluation

Experiment Setup: On an Intel Core 2, 2.33GHz, with 2GB of memory

Evaluation Method : No data available now.

Hypothesis: No data available now.

Description: No data available now.

Dimensions: No data available now.

Benchmark used: No data available now.

Results: No data available now.