Abstract

The dependences between structure and content of XML document cannot be effectively captured by the traditional tree-based modeling method, and thus leads to the XML keyword query difficult to achieve a high recall and precision. Using the semantic analysis techniques, this paper proposes a novel XML document modeling and keyword query approach, which takes full account of the nodes context to optimize the XML document tree modeling and construct the XML Semantic Tree (XST). Based on the XST, the set of Semantically Related (SR) node is firstly defined, and then the rules and algorithm for determining the set of SR nodes are developed. The XML keyword query method, which is based on XST modeling and set of SR nodes, is finally presented. The experimental results demonstrates that the approach proposed in this paper can improve the query precision and recall significantly comparing with the existing typical approaches, and it also has the high performance.

Paper Details