Lexically-generated subject hierarchies for browsing large collections
Developing intuition for the content of a digital collection is difficult. Hierarchies of subject terms allow users to explore the space of topics that a collection covers, to form and specialize useful query terms, and to directly identify interesting documents. We describe two interfaces for navigating such hierarchies, and present a technique for inferring hierarchies automatically from large corpora. We also discuss scalability issues for the techniques involved, and our solutions to these problems.
Unable to display preview. Download preview PDF.