AutoMap4OBDA

Automated generation of R2RML mappings
for Ontology-Based Data Access.

The main motivation for designing this system was to obtain mappings of significantly higher quality compared to the competitors’ results. To do so, the database content and features of the target ontology are taken into account during the matching process. Moreover, AutoMap4OBDA has been designed to be used in OBDA scenarios and is able to generate fully compliant R2RML mappings without user intervention.

AutoMap4OBDA takes as inputs a relational database (i.e., PostgreSQL) and an ontology in OWL to produce a putative ontology from a relational database which is used as an intermediate element in the relational-to-ontology process.

Features

In AutoMap4OBDA, the database content and features of the target ontology are taken into account during the mapping generation process. We have developed three methods that make the mapping process strongly dependent on the input database and the features of the target ontology to increase the performance of the relational-to-ontology mappings.

Ontology learning methods to infer class hierarchies for development of a putative ontology

String similarity metric selection based on target ontology labels for ontology alignment

Short path strategy for R2RML mapping generation based on alignments

Download

AutoMap4OBDA repository will be released soon

Publications

Á. Sicilia and G. Nemirovski "AutoMap4OBDA: Automated Generation of R2RML Mappings for OBDA", To appear in 20th International Conference on Knowledge Engineering and Knowledge Management (EKAW). Bologna, Italy, 19-23 November 2016.

Contact

ARC Enginyeria i Arquitectura La Salle - Universitat Ramon Llull
Álvaro Sicilia (asicilia@salleurl.edu).