Vector retrieval model for XML document based on dynamic partition of information units

XML document is applied in WEB application more and more. Because users can find what they need in numerous XML documents, technology of information retrieval based on XML document becomes a hot topic in information retrieval field now. Traditional technology of information retrieval based on XML document need define retrieval unit and retrieval result unit of the retrieval beforehand, and the dividing granularity is either too big or too small. In this paper we propose a retrieval method, which can dynamically partition information units in terms of the structure and semantic information of XML in vector space model. Therefore it reduces calculating workload efficiently and improves running efficiency of the entire retrieval system. The retrieval efficiency of this method is proved than the traditional one when they have the same accuracy. Finally, the results have been testified by experiment.