Word sense disambiguation is to determine a proper sense in a context for a word which has more than two senses. The task is very important for natural language processing systems such as machine translation systems and information retrieval systems t... Word sense disambiguation is to determine a proper sense in a context for a word which has more than two senses. The task is very important for natural language processing systems such as machine translation systems and information retrieval systems to improve their performance. Therefore, it has been studied widely in various ways such as a method using dictionary, a supervised learning method using sense tagged corpus, an unsupervised learning method using raw corpus, ect. Word sense disambiguation for Korean language has been studied in various ways too, including a method using sense tagged corpus based on word space model which showed relatively good performance. However, the method has weak points such as too high dimensions as the learning words increase and no solution for unknown words. In this , I propose embedded word space models for word sense disambiguation using word embedding to reduce the dimension. The models are proposed with 4 different ones combining space elements: word or sense, and word embedding architectures: CBOW (continuous bag of word) or skip gram. The experiment on Sejong morpheme sense tagged corpus showed that the best model performed better than previous model in spite of reduced dimension of word space model. ，韩语论文题目，韩语毕业论文 |
