Graph-based clustering for semantic classification of onomatopoetic words

Abstract

This paper presents a method for semantic classication of onomatopoetic words like "ひゅーひゅー (hum)" and "からん ころん (clip clop)" which exist in every language, especially Japanese being rich in onomatopoetic words. We used a graph-based clustering algorithm called N ewman clustering. The algorithm calculates a simple quality function to test whether a particular division is meaningful. The quality function is calculated based on the weights of edges between nodes. We combined two different similarity measures, distributional similarity, and orthographic similarity to calculate weights. The results obtained by using the Web data showed a 9.0% improvement over the baseline single distributional similarity measure.

Features of Distributional Method for Indonesian Word Clustering

“…A number of semantic clustering algorithms have been reported, such as those in [8,[10][11][12][13][14][15][16][17][18]. Some work has thus focused on a re-ranking strategy, Geffet and Dagan [12,19] improved the output of a distributional similarity system for an entailment task using a web-based feature inclusion check, and comment that their filtering produces better outputs than cutting off the similarity pairs with the lowest ranking.…”

Use of Sound Symbolism in Sentiment Classification

“…For example, (Komiya and Kotani 2011) proposed a method for clustering Japanese onomatopoeic words by using single-link hierarchical clustering based on context. (Ichioka and Fukumoto 2008) proposed a method for clustering Japanese onomatopoeic words on the basis of the co-occurrence of onomatopoeic words on the Web and the vocal sound similarity of the words. However, the vocal sounds they considered were insufficient to represent pronunciation.…”