AGODW的参考文献 Bootstrapping Domain Ontology for Semant

Semantic Web Services from Source Web Sites Wensheng Wu1, AnHai Doan1, Clement Yu2, and Weiyi Meng3 1 University of Illinois, Urbana, USA 2 University of Illinois, Chicago, USA 3 Binghamton University, Binghamton, USA Abstract. The vision of Semantic Web services promises a network of interoperable Web services over dierent sources. A major challenge to the realization of this vision is the lack of automated means of acquiring domain ontologies necessary for marking up the Web services. In this paper, we propose the DeepMiner system which learns domain ontologies from the source Web sites. Given a set of sources in a domain of inter- est, DeepMiner rst learns a base ontology from their query interfaces. It then grows the current ontology by probing the sources and discov- ering additional concepts and instances from the data pages retrieved from the sources. We have evaluated DeepMiner in several real-world do- mains. Preliminary results indicate that DeepMiner discovers concepts and instances with high accuracy. 1 Introduction Past few years have seen an increasingly widespread deployment of Web services in the e-commerce marketplace such as travel reservation, book selling, and car sale services [21]. Among the most prominent contributing factors are several XML-based standards, such as WSDL [26], SOAP [20], and UDDI [22], which greatly facilitate the specication, invocation, and discovery of Web services. Nevertheless, the interoperability of Web services remains a grand challenge. A key issue in enabling automatic interoperation among Web services is to semantically mark up the services with shared ontologies. These ontologies typi- cally fall into two categories: service ontology and domain ontology. The service ontology provides generic framework and language constructs for describing the modeling aspects of Web services, including process management, complex ser- vice composition, and security enforcement. Some well-known eorts are OWL-S [5],WSFL [11], an