A Prufer Sequence Based Approach to Measure Structural Similarity of XML Documents

Abstract

XML is a W3C standard for exchange of semi-structured data. For many applications it is necessary to extract information from semi-structured data which is a complex task. In this paper we address the problem of computing structural similarity of XML documents which play a crucial role in clustering process. Previous works on path based approach fail to capture the sibling relationship among the nodes and also ignore the similarity when the nodes in the paths to be matched, are not in the same order but still convey same semantics. Another weakness of this approach is in the case of the partial path match, is that the level information is not taken into account when the nodes to be compared appear in different hierarchical level. To address these issues, we describe a method based on Prufer Sequence for measuring the structural similarity of XML documents, in this paper. Benefit of Prufer sequence based representation is that, it stores the ancestor-descendant and sibling relation. XML trees are encoded based on Prufer sequence which establishes a one-to-one mapping between XML tree and sequence. Instead of extracting all paths only common nodes are extracted based on Prufer sequence code. We have devised an algorithm to compute similarity by exploring all relations among the common nodes namely parent-child, ancestor-descendant and sibling. The experimental results show that the proposed approach is effective.