| 研究実績の概要 |
During the last year, I showed through a case study that currently available language learning curriculum have does not answer the real-life needs of language learners as the words taught cover a very small part of the words language learners need. Consequently, I proposed to take advantage of today's availability of learner data and proposed a framework for personalized vocabulary learning. The framework consists of two main parts. The first part provides language learners translation based on their learning logs. The second part is a demographic based, purpose-based and content-based recommendation system. The recommendation system suggests vocabulary that other people with the same demographics or purpose searched for in their digital dictionaries. Learners' data also enables content-based recommendations by suggesting vocabulary that is thematically similar to the one translated by the learner in the past. Additionally, the past logs are used to detect a change in purpose or recommend words based on the situated context of the learner.
We tested each part of the system separately, and showed that the personalized translation can detect the intended meaning of the learner better than Google Translate. We also showed that recommending vocabulary to language learners using their past learning activity increases their learning achievement and motivation.
|
