Improving Extraction of Japanese Functional Expressions with Discontinuous Types through Part-Of-Speech Tagging
Jun Liu, Renting Chen, Chenrui Liang · 2023
In learning process of Japanese as a second language (JSL), the acquisition of grammatical knowledge is the requisite foundation to learn Japanese well. There are various types of functional expressions in Japanese grammar, which is one of the difficult problems in Japanese language teaching and learning. This paper proposes a simple but effective method to Japanese language processing by Part-Of-Speech (POS) tagging with morphological analysis technology, which handles the automatic extraction of the discontinuous Japanese functional expressions for assisting JSL learners with their study of Japanese grammar. In this task, discontinuous functional expressions in Japanese example sentences will be automatically extracted based on the output results generated by a well-known Japanese morphological analysis tool, named MeCab. To accomplish this task, we manually annotated the POS information for the discontinuous Japanese functional expressions to train a new Conditional Random Fields (CRF) model, which can be applied to Japanese morphological analysis. The evaluation results revealed that our method works effectively on automatic of the Japanese discontinuous functional expressions. The method can also be applied to the development and utilization of new effective functions in computer-assisted language learning system (CALL) for Japanese language teaching and learning.