K-012
Investigations of Difficulty Level Update in Recommendation Function for Element Fill-in-Blank Problem in Java Programming Learning Assistant System

◎Su Sandy Wint・Nobuo Funabiki・Minoru Kuribayashi（Okayama Univ.）

To benefit Java programming studies, we have studied a Web-based Java programming Learning Assistant System (JPLAS)that offers the element fill-in-blank problem (EFP)for a novice student to study Java grammar and basic programming skills. By applying the blank element selection algorithm, a lot of EFP　instances have been generated to cover the whole grammar study. Previously, we proposed the recommendation function　to suggest a proper EFP　instance for each student to be solved next. It considers the difficulty level　and the solving result of the previous instance. This difficulty level is updated in the linear curve to the number of solved instances.In this paper, we investigate the effect of the update curve into the student satisfaction. Our preliminary experiment results show that the quick start curve gives the higher satisfaction than the liner or slow start ones.