Automatic Document Clustering using Topic Analysis
Robert J. Muscat · OAR@UM (University of Malta) · 2004
Abstract: In this work we look at the automated organisation of documents in a document base into clusters and cluster hierarchies. We apply topic segmentation to detect topics within documents and using term relationships attempt to build hierarchies which represent a “real world ” topic hierarchy. Finally, we assign documents to each of these topics using a standard clustering technique. We also propose two evaluation methods for document clustering systems. The first is an adaptation of tree measure algorithms to document hierarchies. The use of this method will require a pre-defined tree which has been agreed upon as a suitable benchmark. The second is independent of any benchmark trees and presents the evaluator with a number of measures which allow him to assess the properties of the tree. Automatic Document Clustering using