Measuring Similarity between XML Documents

With the advance of World Wide Web standards, XML documents become popular in e-business applications for information exchange. Electronic catalogs and transaction records are now formatted in XML. XML documents are semi-structured documents with XML schemas marking up the semantics. XML separates presentation from semantics so that presentation of information on different devices can be processed independently from information management. However, information retrieval techniques for XML documents are still in the beginning stage. Most ebusiness applications only use XML for data interchange or presentation purposes. Advance information retrieval techniques have not been applied in retrieving, organizing, and managing XML documents. In this paper, we propose a similarity measurement between XML documents based on the Jaccard’s similarity function for unstructured text documents. Examples will be provided to illustrate the formulation. Given the proposed similarity measurement, traditional information retrieval techniques for unstructured test documents can also be applied for XML documents. The performance of XML document management will be significantly improved.