Discovering frequently changing structures from historical structural deltas of unordered XML

Abstract

Recently, a large amount of work has been done in XML data mining. However, we observed that most of the existing works focus on the snapshot XML data, while XML data is dynamic in real applications. To the best of our knowledge, none of the existing works has addressed the issue of mining the history of changes to XML documents. Such mining results can be useful in many applications such as XML change detection, XML indexing, association rule mining, and classification etc. In this paper, we propose a novel approach to discover the frequently changing structures from the sequence of historical structural deltas of unordered XML. To make the structure discovering process efficient, an expressive and compact data model, Historical-Document Object Model (H-DOM), is proposed. Using this model, two basic algorithms, which can discover all the frequently changing structures with only two scans of the XML sequence, are presented. Experimental results show that our algorithms, together with the optimization techniques, are efficient and scalable. Copyright 2004 ACM.