Harvesting Wiki Consensus - Using Wikipedia Entries as Ontology Elements

IUFRO 7 Division Forest Health Project: "Forests under Changing Climate, enhanced UV and Air Pollution"

Martin Hepp ; Daniel Bachlechner ; Katharina Siorpaes

Others  :  http://CEUR-WS.org/Vol-206/paper10.pdf
PID  :  528

来源: CEUR

【 摘 要 】

One major obstacle towards adding machine-readable annotation to existing Web content is the lack of domain ontologies. While FOAF and DublinCore are popular means for expressing relationships between Web resourcesand between Web resources and literal values, we widely lack unique identifiersfor common concepts and instances. Also, most available ontologies have avery weak community grounding in the sense that they are designed by singleindividuals or small groups of individuals, while the majority of potential usersis not involved in the process of proposing new ontology elements or achievingconsensus. This is in sharp contrast to natural language where the evolution ofthe vocabulary is under the control of the user community. At the same time,we can observe that, within Wiki communities, especially Wikipedia, a largenumber of users is able to create comprehensive domain representations in thesense of unique, machine-feasible, identifiers and concept definitions which aresufficient for humans to grasp the intension of the concepts. The Englishversion of Wikipedia contains now more than one million entries and thus thesame amount of URIs plus a human-readable description. While this collectionis on the lower end of ontology expressiveness, it is likely the largest livingontology that is available today. In this paper, we (1) show that standard Wikitechnology can be easily used as an ontology development environment fornamed classes, reducing entry barriers for the participation of users in thecreation and maintenance of lightweight ontologies, (2) prove that the URIs ofWikipedia entries are surprisingly reliable identifiers for ontology concepts, and(3) demonstrate the applicability of our approach in a use case.