利用自然語言處理技術建構網路日語文法輔助學習系統──以形態素解析JUMAN、構文解析KNP之應用為例

A Japanese Grammar Learning Support System Using Natural Language Processing Technology-The Case of a Morphological Analysis System JUMAN and a Sentence Structure Analysis System KNP

本研究以日本語能力檢定N2程度之文法知識習得為目標，開發出線上輔助學習系統。研究中使用了「新日語検攻略問題集：N2文法」作為授課教材。為了突破以往文法學習的限制以及學習效率低落等問題，藉由結合教室授課與線上自主學習，分成上課前、教室上課與上課後三階段進行文法指導與學習。首先在預習的階段，由學生利用系統中的KNP、JUMAN等自然語言處理技術，即時分析輸入之學習內容，並針對系統產生之較容易了解之句子結構與詞彙資訊進行預習。此外再透過多國語言辭典與相關文法表現等功能學習，以求更深入了解內容。教室授課部份，學生則接受教師的文法學習指導，確認之前預習內容，並藉由聆聽教師講解，以釐清在預習階段時無法理解之內容。最後在進行下一階段的新的內容學習前，從系統的學習履歷中，檢索出之前預習的內容，再透過合成語音與語音辨識功能，重新以聆聽的方式進行複習。與以往教室文法授課方式比較之實驗評估結果顯示，除了確認本研究之有效性外，也發現學生感受到自身的聽力和口語表達能力有所提升，對這樣的學習方式感到滿意。

In this study, a web-based learning support system was developed to facilitate the acquisition of JLPT N2 level grammatical knowledge. To overcome the constraints of grammar learning and solve the problems of inefficiency, instruction was compiled in the form of integration of in-class instruction and online autonomous learning that was carried out in three phases ''preview'', ''in-class instruction'' and ''review''.At first learners were asked to use KNP and JUMAN to analyze and preview learning contents, then use multi-language dictionary and grammatical expression function to verify their actual usages. The in-class instruction phase plays a role of clarification for preview learning results under the direction of instructor. Finally, learners retrieve their learning records from system and use voice synthesis and speech recognition functions to review the contents.An evaluation experiment was conducted. The results indicate that the proposed system was effective in improving learners' knowledge in grammar as well as their ability in listening and speaking. Learners also expressed their high level of satisfaction toward the proposed way of grammar learning.