A Method For Encoding XML-based Tree Structures And Measuring Structural Similarity

With the rapid development of Internet and accordingly the data, the requirements to webinformation are increasing. But the traditional Web information inquiry majority is based onthe HTML language, because HTML in the semantics expressed in the aspect has the certainflaw, has caused XML and the correlation technology fast development. And accordingly howto search information rapidly, efficiently and generally based on XML , and further excavate thestructure and the semantic information, has become an urgent issue. This paper proposes a codingapproach for tree structure based on XML to solve this problem. It adopts coding to inquiredata source and uses structural similarity to match inquiry results, and then provide theinformation according to users' requirements. This paper will pay attention to such aspects:First, This article introduced the XML retrieval technology development presentsituation, has carried on the review to information retrieval technology each method,summarized to has carried on the match based on the XML documents tree structure the basicprinciple and the classification.Second, this paper proposes an inquiry technique .The technique first makes index labels(that is to coding) on the basis of tree structure for XML file. Then it separates out the nodessatisfying the match rules by inquiring nodes in order to contain the structure relationshipamong nodes. Finally it calculates the match degree according to the match calculationformula given in this paper and offers users the results they needed.Then, this article through gave an example to analyze showed this method characteristic:1. It reduces the data volume for re-scanning when the data source changing.2. in the inquiry process, may make the user clearly to understand each step of point information theinquiry situation.3. It can guarantee the certain the accuracy and the completeness for the structuralsimilarity matches are made based on inquiring effective nodes, that is to makematches ensuring users' requirements.Finally, this paper also discusses such problems as the trees' expand, subtrees' inquiryand trees' pruning to seek ways to implement intelligent information index.