Architecture for Mapping Relational Database to OWL Ontology: An Approach to Enrich Ontology Terminology Validated with Mutation Test Cristiane A. G. Huve 1,2 a , Alex M. Porn 1 b and Leticia M. Peres 1 c 1 Department of Informatics, Federal University of Paran´ a, Av. Cel. Francisco H. dos Santos, 100, Curitiba, Brazil 2 Polytechnic School,Uninter, Rua Luiz Xavier, 103, Curitiba, Brazil Keywords: Mapping, Ontology, Relational Database, Mutation Test. Abstract: Ontologies are structures used to represent a specific domain. One well-known method to simplify the on- tology building is to extract domain concepts from a relational database. This article presents an architecture which enables an automatic mapping process from a relational database to OWL ontology. It proposes to en- rich the terminology of ontology elements and it was validated with mutation tests. The architecture mapping process makes use of new and existent mapping rules and overcome lacks not previously addressed, such as the use of database logic model to eliminate duplicated elements of ontology and mapping inheritance rela- tionships from tables and records. We stand out the structure of element mapping, which allows maintaining source-to-target traceability for verification. We validate our approach with two experiments: the first one fo- cuses on architecture validation applying an experiment with three scenarios and the second one uses a testing engine applying a mutation test methodology to OWL ontology validation. 1 INTRODUCTION In computing, an ontology is defined based on a set of concepts in which a domain of specific knowledge is modeled (Gruber, 1995). Ontologies offer advantages in their use, such as: providing an exact description and an exact vocabulary for representation and shar- ing of knowledge (Guarino, 1995). The process of its elaborating is a task which requires a great amount of effort (Staab and Studer, 2013; Telnarova, 2010). Several approaches aim to convert relational databases into ontologies. Most parts of these so- lutions develop a mapping process from a set of rules that considering a relational database (RDB) and physical model. Michel et al. (2014), Spanos et al. (2012) and Sequeda et al. (2011) surveyed the motivations and the benefits of a mapping process from relational database to ontology, considering the challenges, and different application purposes. Al- though some authors use the R2RML language (Se- queda et al., 2012; Das et al., 2012; Arenas et al., 2012) and it is an important advance for the commu- nity, the R2RML language has restrictions and does a https://orcid.org/0000-0002-2038-9450 b https://orcid.org/0000-0003-0832-5750 c https://orcid.org/0000-0002-8922-6975 not support to record during the mapping the relation- ships between RDB and ontology concepts. Concerning the creation of an ontology from scratch since an RDB, we analyzed related work as Astrova (2009); B¯ umans and ˇ Cer¯ ans (2010); Cullot et al. (2007); Gherabi et al. (2012); Laclavik (2006); Li et al. (2005); Louhdi et al. (2013); Ramathilagam and Valarmathi (2013); Ren et al. (2012); Telnarova (2010); Vavliakis et al. (2010); Zhang and Li (2011) and Jain and Singh (2013). Tissot et al. (2019) de- scribes in detail such related work mentioned above, being explained rules commonly used in the map- ping process. Huve (2017) explores related work and developed experiments with Astrova (2009) work, which has made a considerable contribution to this re- search area. Astrova (2009) proposed QUALEG DB, a tool to deal with hierarchy, constraints, and restric- tions of elements, however, it works with direct map- ping. More recently, Jain and Singh (2013) compared different approaches developed to convert RDB to on- tology, and they proposed adding new features to an existing framework. Nevertheless, this proposal pro- vides a direct mapping between RDB and ontology elements, limiting the possibilities of ontology mod- eling. In general, these works do not expose details about the terminology of naming ontological el- 320 Huve, C., Porn, A. and Peres, L. Architecture for Mapping Relational Database to OWL Ontology: An Approach to Enrich Ontology Terminology Validated with Mutation Test. DOI: 10.5220/0007752803200327 In Proceedings of the 21st International Conference on Enterprise Information Systems (ICEIS 2019), pages 320-327 ISBN: 978-989-758-372-8 Copyright c  2019 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved