Data Integration In A Number Of Issues

How to deal with non-structural text data effectively has become a hot and hard research topic in data integration domain.In this paper,we made two efforts in accordance with this issue.We proposed a uniform semantics-based data integration model,i.e.the ontology-tree model, which is domain-independent,can be used to solve problems of data integration from different deep web data domains(such as books,autos, etc.).This model can provide data retrieval services based on data semantic relationships.Experiments demonstrated that the model is efficient and with highly flexibility in practice.Meanwhile,we improved the inter-relevant successive tree model,an existed full-text index model, in encoding aspect.We proposed a bit-based encoding solution,which reduces the expansion rate dramatically.And,the related text-regenerating algorithm and text-retrieval algorithm have got good results as well.In addition,we analysed the application problem of XML data,which has become one of the critical tools today in data integration domain.We proposed the Value Object-Oriented Model for XML data, which settles two outstanding hard issues in XML data's real applications.