ONLINE LEGAL-AID - MARKOV-CHAIN MODEL FOR EFFICIENT RETRIEVAL OF LEGAL DOCUMENTS

It is widely accepted that, with large databases, the key to good perf ormance is effective data-clustering. In any large document database c lustering is essential for efficient search, browse and therefore retr ieval. Cluster analysis allows the identification of groups, or cluste rs, of similar objects in multi-dimensional space [1]. Conventional do cument retrieval systems involve the matching of a query against indiv idual documents, whereas a clustered search compares a query with clus ters of documents, thereby achieving efficient retrieval. In most docu ment databases, periodic updating of clusters is required due to the d ynamic nature of a database. Experimental evidence, however, shows tha t clustered searches are substantially less effective than conventiona l searches of corresponding non-clustered documents. In this paper, we investigate the present clustering criteria and its drawbacks. We pro pose a new approach to clustering and justify the reasons why this new approach should be tested and (if proved beneficial) adopted. (C) 199 8 Elsevier Science B.V. All rights reserved.