Abstract
The study applies graph theory and network analysis methods to investigating the characteristics of a semantic network representation of the Japanese Word Association Database (Joyce, 2005, 2006). After briefly outlining the Japanese Word Association Database, this study describes the application to the word association network of graph clustering methods, particularly a recently proposed graph clustering algorithm called RMCL (Jung, Miyake, & Akama, 2006a, 2006b). Analysis results indicate that the developed network has scale-free characteristics. Comparisons of clustering techniques demonstrate the usefulness of the RMCL method, and its potential as a tool for visualizing large-scale linguistic resources, such as the Japanese word association database.
