XML documents are semistructured and the structure of the documents is embedded in the tags. Although XML doc-uments can be accompanied by a DTD that defines the struc-ture of the documents, the presence of a DTD is not manda-tory. The difficulty in deriving the DTD for XML documentslies in the fact that DTDs are of different syntax as XMLand that prior knowledge of the structure of the documentsis required. In this paper, we introduce DTD-Miner, an au-tomatic structure mining tool for XML documents. Using aWeb-based interface, the user will be able to submit a setof similarly structured XML documents and the system willautomatically suggest a DTD. The user is also able to fur-ther refine the DTD generated to reduce the complexity byrelaxing some the rules used in the system...

The emergence of the Extensible Markup Language (XML) [7] provides a partial solution to this problem. XMLis a structured format for data interchange over the Web.The main difference between HTML and XML (and also the reason for the uprising of XML over HTML) is theuse of tags - while HTML tags are primarily used to de-scribe they way in which a data item is displayed on Webbrowsers, XML tags describe the data itself. The effect isthat XML documents are self-describing and this facilitatesthe post-processing of Web documents, promising a com-plete change in the “behavior” of the Web in the near future.