Paper: Topic Identification Using Wikipedia Graph Centrality

ACL ID N09-2030
Title Topic Identification Using Wikipedia Graph Centrality
Venue Human Language Technologies
Session Short Paper
Year 2009
Authors
  • Kino Coursey (University of North Texas, Denton TX; Daxtron Laboratories, Inc., TX)
  • Rada Mihalcea (University of North Texas, Denton TX)

This paper presents a method for automatic topic identification using a graph-centrality al- gorithm applied to an encyclopedic graph de- rived from Wikipedia. When tested on a data set with manually assigned topics, the system is found to significantly improve over a sim- pler baseline that does not make use of the ex- ternal encyclopedic knowledge.