Jump To: Parent Description
- GWSD is a system for Unsupervised Graph-based all-Words Word Sense Disambiguation. Please refer to (Sinha and Mihalcea, 2007) for a description of the graph-based disambiguation method, as well as for brief descriptions of all the similarity measures and the graph-centrality algorithms used by GWSD. For a quick trial of GWSD, you can use some of the pre-built graphs and feature files provided with the distribution. These graphs are stored in folders whose names clearly specify the type of the graphs (i.e. the corpus, window size, part-of-speech used, etc.). One example of such a set of graphs stored for, say, Senseval-2, a window size of 2, part of speech 'noun', and a similarity measure 'jcn', would be as follows: The set of graphs, one graph for each word to be disambiguated, will be located inside the folder 'Senseval-2.jcn.n.2.Graphs'.