Abstract:

A word sense disambiguation method based on semantic graph structure was presented. In the method, a text is represented as an undirected weighted semantic graph with WordNet, which defines synsets as vertices and relations of vertices as edges, and assigns the weight of edges with the relatedness of connected synsets. Combining with coherence synset and the synset frequency, PageRank* which is a modified PageRank formula for undirected weighted graph, was used to disambiguate the nouns in the text. The experimental result shows that this method is effective and practical.

Key words: word sense disambiguation, semantic graph, WordNet, PageRank