Exploring Large Document Collections with a Dynamic Hierarchy
Alexander F. Gelbukh, Grigori Sidorov · 2004
We present a method for dealing with the large document collections that takes into account the user needs in a better way. The user can choose the desired granularity of the hierarchy and then apply this hierarchy to document collection for classifying the documents. The granularity depends directly on the number of clusters. The clusters are presented to the user in two different ways: (1) as a representative document of the cluster; (2) as a set of keywords characterizing all documents of the given cluster.