Abstract—Computer-based text analyses have been widely applied to the Japanese language. However, the application is complicated because a word order is flexible in Japanese sentences; thus, the role of nouns cannot be determined only by sentence structure. Such roles are referred to deep cases that represent the essential meaning of nouns, which is important in the semantic analysis of sentences. Identifying deep cases of nouns can improve text analysis results. Since surrounding words are important to determine these roles, their identification should have high affinity to word embedding, which has gained popularity in natural language processing. In word embedding, similar role words have similar vectors with high cosine similarity. Therefore, in this paper, we propose a semantic analysis method using word embedding.
Index Terms—Deep cases, semantic role labeling, semantic analysis.
Takumi Kawasaki is with the Electrical Engineering and Computer Science Department, Shibaura Institute of Technology, Tokyo Japan (e-mail: ma17031@shibaura-it.ac.jp). Masaomi Kimura is with the Information Science and Engineering Department, Shibaura Institute of Technology, Tokyo Japan (e-mail: masaomi@shibaura-it.ac.jp).
[PDF]
Cite:Takumi Kawasaki and Masaomi Kimura, "Deep Case Identification Using Word Embedding," International Journal of Computer Theory and Engineering vol. 10, no. 6, pp. 216-220, 2018.
