XML data is an acknowledged typical semi-structured data with grammar and without semantic information. However sematic information is very important to preserve data integrity, data storage and query optimization. Based on the existing XML data management motheds, this paper studies some key issues related to XML data management techniques. In order to maximize semantic power of XML specifications, the main ideas on integrity constraints for XML documents and the problems related to them are described and discussed. Then we proposel a Contrainted XML Data Model (CXM), and make some extensions and standardizations on XML-QL. Finally, an XML Data Management System (CXDMS) is implemented. Experimental result shows that the methods put forward in this paper can efficiently store and manage XML data. Additionally, we design and implement a data-mining tool based on artificial neural networks to exchange rate forecasting and acquire better experimental effect. |