Classifying Japanese Polysemous Verbs based on Fuzzy C-means Clustering

This paper presents a method for classifying Japanese polysemous verbs using an algorithm to identify overlapping nodes with more than one cluster. The algorithm is a graph-based unsupervised clustering algorithm, which combines a generalized modularity function, spectral mapping, and fuzzy clustering technique. The modularity function for measuring cluster structure is calculated based on the frequency distributions over verb frames with selectional preferences. Evaluations are made on two sets of verbs including polysemies.