Indexing XML Documents: Improving the BUS Method Vincent Oria, Amit Shah, and Samuel Sowell   View Paper (PDF)   Return to Papers Abstract In this paper we propose an indexing method for XML documents that improves the BUS method. The main difficulty in indexing XML documents is that the index must accommodate two different worlds that XML brings together: database and information retrieval. In the database world, indexing is based on the data structure, while indexing in information retrieval is based on the text content. One of the original XML document indexing methods is the Bottom Up Scheme (BUS). This method was then modified to create a better mechanism for updating structured documents. Despite the significant improvements made to the Bottom Up Scheme, the experiment still indicated a 240% storage overhead, indexing time of 4 hours and 43 minutes for 250 MB of documents, and query times as long as 6.46 seconds. The solution to minimizing the overhead introduced by the indexing is to reduce the number of terms to be indexed. The method proposed in this paper selects important words from the documents to be indexed, based on a user-defined dictionary. Furthermore, each node in the tree structure of the BUS method has a fixed number of children that causes complete tree reorganization from time to time. We propose a solution that avoids the reorganization of tree. DiSC'02 © 2003 Association for Computing Machinery