Data Classification and Question Generation in OSARU
"A web based Japanese language acquisition support system"



"OSARU", a web based Japanese language acquisition support system, is a Computer Aided Language Learning (CALL) system that has been developed at the Computational Linguistics Laboratory since 2001. The system's goal is to create an environment for effective teaching and learning of Japanese language. It semi-automatically generates language exercises that are tailor-made to each user to save time and labor of teachers.

To generate the tailor made questions, the system needs a corpus and the database with the characters and vocabulary. Texts, characters and words should be carefully classified according to different aspects so that the system's interface can process it appropriately.

This thesis focuses on two parts of the OSARU system. First one is the classification of the characters and vocabulary with respect to difficulty and fields they belong to and the database system created for their storage and use. The second part discusses the decoy generation procedure for kanji selecting questions, and the analysis of Japanese grammar, in which onomatopoeia and mimesis are discussed.


Computer Aided Language Learning, Japanese Language Proficiency Test, Difficulty Value, Field Value, Relative Frequency, Decoys, Onomatopoeia, Mimesis