Iterative Ontology Alignment Debugging Using a Scenario- and Strategy-Driven Approach

The automatic matching of entities between information repositories is essential for aligning and ontology mediation processes, but there is great ambiguity when adopting it. Data integration between repositories demands great quality of ontology alignments, and as such ambiguous correspondences must be identified and corrected beforehand. Debugging is even more relevant if the process is systematic and complete, thus allowing its formalisation and implementation and promoting its acceptance. In this paper we address the analysis and systematization of the ontology alignment debugging process, proposing the characterization of the ontology matching scenarios through ten dimensions. The resulting scenarios are categorized as Simple and Composed depending on the associated actions taken for solving the ambiguity. We suggest evolving the alignment process into an iterative matching-debugging process, where the resulting ambiguous scenarios and correction strategies are fed into the matching algorithms for supporting the resolution of the alignment problems.