One Improved Description Method Of Data Sources In The Deep Web

As the development of the web, we witness a rapid increase in the number of datasources that are available in deep web. These data sources include commercial databases onproduct information, stock market information, real estate, automobiles, books andentertainment. They have many characteristics such as large amount、professional、highvaluable、more hidden and heterogeneous, etc. It is an important research area to integrate thedeep web for the full use of these data sources and the help of providing interesting and usefulinformation for users.Source description plays such an important role in the deep web integration system that itis not only the basic of source queries in the deep web, and it has close relationship with theinterface integration and pattern matching in the deep web. At present, there exist two maindirection in source description: the first one is based on a global view through which to createmapping relationships to the data sources, this way has the very good scalability that addingor deleting a data sources will not affect the global view, but it is difficult to build a goodglobal view; the other is based on the template description and using a specific algorithm tocompute the mediator capabilities, this way can improve the efficiency of the source inquires,but adding or deleting a data source may affect the mediator view.In this paper, the main contributions represent in the following several aspects:(1) theclassification and related definitions of the form attributes in the deep web have been donefirstly.(2) The two typical deep web integrating systems are compared.(3) The method ofsource description are compared, and a new method of source description is proposed whichis different from the on based on the adornments template but is based on the bitmap template,this method has expands the capability description of data sources and can support interfaceintegration.(4) Based on the source ability, this paper has put forward a method of computingthe mediator capability.