The Preliminary Process of Modeling in Deep Web Information Fusion System

The Web has become the largest information source, which includes all aspect of human´s life. The myriad information is believed to be hidden behind the deep Web, which the search engines and crawlers can´t access directly. To extend the human´s physical limitation of accessing information, the information fusion system is introduced. The Web is very different from the traditional database community, the nature of which is more dynamic, heterogeneous, dispersed and hyper-linked. The query interfaces of the deep Web are the only clue to disclose the hidden schemas. However, to get schemas of the local databases through the query interfaces is a great challenge, which is deserved dramatically concern. The key aspect of the deep Web information fusion is the source modeling problem, which is about the specification of the correspondence between the data at the sources and those in the global schema. The first step is to parse the query interfaces and extracting attributes and semantic relationships which reflect the hidden schemas. There are two approaches of source modeling: global as view (GAV) and local as view (LAV). In this paper, we choose the method of LAV and present an attributes extraction method utilizing ontology.