Comparison of optimal path finding techniques for minimal diagnosis in mapping repair

Ontology matching produce a set of semantic correspondences called alignment. The issue of incoherent alignment has been the concern of many researcher since 2010, since almost all matching systems produce incoherent alignments of ontologies. Mapping repair process is a way to quantify the quality of alignment based on the definition of mapping incoherence. Internal properties of mapping will be measured by semantic of the ontologies being matched. Mapping repair process should restore coherence condition by removing as less as possible unwanted mappings. This is call minimal diagnosis. Minimal on the amount of removed mapping and small confidence value of removed mapping. This paper compares optimal path finding techniques that support minimal diagnosis. Some experiments conducted using conference track ontology. Experiment result showed that A∗ Search produced the greatest precision, recall and f-measure values, followed by Greedy Search. Both techniques computed the lowest cost path by using heuristic. This condition was also due to logic algorithm that effective to support minimal diagnosis.