Collaborative filtering for expansion of learner’s background knowledge in online language learning: does “top-down” processing improve vocabulary proficiency?

In recent years, collaborative filtering, a recommendation algorithm that incorporates a user’s data such as interest, has received worldwide attention as an advanced learning support system. However, accurate recommendations along with a user’s interest cannot be ideal as an effective learning environment. This study aims to develop and evaluate an online English vocabulary learning system using collaborative filtering that allows learners to learn English vocabulary while expanding their interests. The online learning environment recommends English news articles using information obtained from other users with similar interests. The learner then studies these recommended articles as a method of learning English. The results of a two-month experiment that compared this system to an earlier collaborative filtering system called “GroupLens” reveal that learners who used the collaborative filtering system developed in this study read various news articles and had significantly higher scores on topic-specific vocabulary tests than did those who used the previous system.