Word Sense Disambiguation based on Relation Structure

摘要

This research is to disambiguate word sense occurred in document. We have studied a research to create a lexical dictionary automatically for unknown word that is generated newly by social trend, new technology, new products, and so on. The mandatory thing of the processes is word sense disambiguation (WSD). Many researches for WSD have showed good results however, they have a limitation to adapt for automation of our final objective. Therefore, we propose new WSD method. For WSD, we design a relation structure (RS) that measures relationship and distance between terms using WordNet. Based on the RS, we could grasp an exact sense of word. Through an experiment and evaluation, our method is as wonderful as a structural semantic interconnection (SSI) that is the best research for WSD.