Given the rapid growth and snccess of publishing information sources on the Web, it is increasingly amactive to wrapper information from distributed, autonomous and heterogeneous Web data sources. Wrapping structured data from web is not a trivial task. Most of the information on the web today is in the form of Hypertext Markup Language (HTML) documents, which are viewed by a browser. HTML documents are sometime written by hand, sometimes with the aid of HTML tools. Given that the form of HTML documents is designed for presentation purposes, not automated extraction. Some Web content is being available in formats more suitable for automated processing, in particular the Extensible Markup Language (XML)[Z]. Despite being a relatively new development, XML has become absolutely essential for enabling data interchange between otherwise incompatible systems. 数据挖掘研究院
Except HTML, XML data sources, there are other types data sources, such as many different datahase data sources and text documents. With the rapid growth of the these types available data sources on the Internet, significant attention has been received on integrating huge heterogeneous information to build a new unified web application. The key technology of these applications is how to homogeneous the difference structures and semantics of the information provided by the enterprise or the Internet. The need to access, retrieves, and manage information from a variety of sources and applications using different data model, representation and interfaces has created a great demand for tools support data and systems integration. Integration of complementary information produced from the popular Web and the rapid growth Intemet, providing useful service for the large amount of the Web users, become a very important and more recent research subject!
We have developed a DIMS mediator-wrapper system to solve this problem. DIMS accepts XML as data model, provides metadata management to facile maintain distributed, autonomous and heterogeneous data sources. DIMS Mediator provides unified XML views of sources containing collections of disparate heterogeneous information that can he queried through the users′ queries. It takes into account the capabilities of the data source, to avoid sending not executable suh-queries to sources. Each data source is encapsulated within a related wrapper. Ideally, a mapper can provide mapping functionalities as XML views to achieve local mappings of data and metadata.
In this paper, we focus on systemoriented issues in Web data wrapper and describe our approach for building a dependable wrapping process. Our ideas are manifested in DIMS, an XMLbased information integration system and have implementing a prototype to demonstrate important properties of the presented approach. 数据挖掘研究院
