Semantic-based Similarity Computation for XML Document

Abstract

XML (eXtensible markup language) is a language for specifying semistructured data, and is rapidly emerging as the new standard for data representation and exchanged on the Web. XML allows users to define elements using arbitrary words instead of a predefined set of words. XML can provide a suggestion for building an index structure for information retrieval or determining a schema for removing data from the Web to a database. However, problems of XML are using different tags name. In this paper, we propose a new methodology for similarity computation between XML documents. The first step is to define tags as the process of extracting tags after making DOM tree, parsing XML document. And, the second step is to make synonym-set as the process of making set by using a glossary defined by users. And, the final step is to measure the similarities among XML documents as the process of measuring similarities among documents with synonym-set and used use-defined main dictionary. When applying the model which presents in this paper, similarity was measured accurately as the time which human compared two documents heuristically.