| Nowadays, as an important platform of network information, Web becomes to be the main resource of getting information. However, because Web pages are non-structure or semi-structure, the data are massive and dynamic, the hyperlinks are disorders, people can hardly get the information they really interested in. There are massive and valuable data and information in the Deep Web. which can not be searched by general search engines, so it will be useful to research techniques of extracting Deep Web information automatically to help people get the knowledge more quickly and more exactly.People proposed the concept of Semantic Web to extract Deep Web information. Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computes and people to work in cooperation. The Semantic Web uses a multi-level framework to achieve its goal. Ontology locates in the level from textual description to knowledge-based reasoning. So it is important to develop ontology for the Semantic Web.Ontology is an explicit specification of a conceptualization. It defines the basic concepts and relations comprising the vocabulary of a topic area. This makes these concepts and relations have explicit and exclusive definitions in certain scope. Then people can communicate with machines freely. Ontology combined with information extracting technology, we use ontology to describe and present the knowledge, it improves the semantic expressing ability of extracting template. We can make the extracting result more veracious by emphasize on one special scope. The information extracting method based on ontology can map word to concept and entity. incarnate the true meaning of the words well, moreover it can embody the relation between the words by class succeed relation of ontology, all this make the information extracting system more powerful.Firstly, give the summary of Deep Web. introduce the definition and types of Deep Web. and quantificationally analyze the overview of Deep Web. Then introduce the definition of Ontology and techniques of DOM, then based on that combine Ontology and techniques of extracting Deep Web information, extract the results of Deep Web pages by comparing sub-tree similarity of web pages. Finally, build a system based on the proposed algorithm. The experiment results showed that we can accurately extract Deep Web result pages in the percent of 85, which proved that the algorithm we presented is effective. |
