Ontologies’ mappings validation and annotation enrichment through tagging

摘要

Pay as you go ontology matching, the technique of first executing an automatic matching tool and then engaging users to improve the quality of an alignment produced by the tool is gaining popularity. Most of the existing techniques employ a single user to validate mappings by annotating them using terms from a controlled set such as “correct” or “incorrect”. This single user based approach of validating mappings using a controlled set of vocabulary is restrictive. First, the use of controlled vocabulary does not maximize the user’s effort since it restrains him/her from adding more meaning to the concepts participating in low-quality mappings using his/her own terms. Secondly, a single user approach of validating a wide range of mappings is error prone since even the most experienced user may not be familiar with all subtopics contained in the input ontologies. We demonstrate in this research that through tagging of concepts participating in mappings flagged as low-quality, we can achieve both mappings’ validation and ontology’s metadata enrichment by adding quality annotations to the ontology.