BootOX : Practical Mapping of RDBs to OWL 2
Ontologies have recently became a popular mechanism to expose relational database (RDBs) due to their ability to describe the domain of data in terms of classes and properties that are clear to domain experts. Ontological terms are related to the schema o
- PDF / 2,072,197 Bytes
- 20 Pages / 439.37 x 666.142 pts Page_size
- 76 Downloads / 165 Views
Department of Computer Science, University of Oxford, Oxford, UK [email protected] 2 Fluid Operations AG, Walldorf, Germany 3 Department of Informatics, University of Oslo, Oslo, Norway 4 Sapienza Universit` a di Roma, Rome, Italy
Abstract. Ontologies have recently became a popular mechanism to expose relational database (RDBs) due to their ability to describe the domain of data in terms of classes and properties that are clear to domain experts. Ontological terms are related to the schema of the underlying databases with the help of mappings, i.e., declarative specifications associating SQL queries to ontological terms. Developing appropriate ontologies and mappings for given RDBs is a challenging and time consuming task. In this work we present BootOX, a system that aims at facilitating ontology and mapping development by their automatic extraction (i.e., bootstrapping) from RDBs, and our experience with the use of BootOX in industrial and research contexts. BootOX has a number of advantages: it allows to control the OWL 2 profile of the output ontologies, bootstrap complex and provenance mappings, which are beyond the W3C direct mapping specification. Moreover, BootOX allows to import pre-existing ontologies via alignment.
1
Introduction
The Semantic Web community has actively investigated the problem of bridging the gap between relational databases and ontologies. One of the main targets behind this effort is to enable the access to the data stored in databases via Semantic Web technologies. An advantage of this approach is that ontologies provide a formal specification of the application domain that is close to the endusers’ view of the domain, while databases are oriented towards an efficient data storage and retrieval and thus represent the data using structures that often not intuitive to end-users. Manually building an ontology and connecting it to the data sources via mappings is, however, a costly process, especially for large and complex databases (a typical scenario in industry [20,22]). The cost of this manual process will typically be even more severe when dealing with multiple databases, e.g., in the context of accessing the Deep Web [16]. To aid this process, tools that can extract a preliminary ontology and mappings from database schemata play a critical role. c Springer International Publishing Switzerland 2015 M. Arenas et al. (Eds.): ISWC 2015, Part II, LNCS 9367, pp. 113–132, 2015. DOI: 10.1007/978-3-319-25010-6 7
114
E. Jim´enez-Ruiz et al.
In the literature one can find a broad range of approaches to bootstrap an ontology and mappings from a relational database schema. The interested reader may have a look at the following surveys [34,41]. These approaches can be classified with respect to different aspects such as:(i) level of automation (i.e., manual, semi-automatic, automatic), (ii) type of mappings (i.e., complex or direct mappings), (iii) language of the bootstrapped mappings and the ontology, (iv) reuse of external vocabularies (e.g., domain ontologies or thesauri), and (v
Data Loading...
