Stepwise Selection of English Multiple-choice Cloze Questions Based on Difficulty-based Features for Keeping Motivation

Abstract

We had constructed the automatic generation system of English multiple-choice cloze questions. By using the system, plenty number of questions can be generated automatically, but learners become difficult to find appropriate questions. Therefore, objective of this research is to develop the method for providing questions that fit for learners. In the self-learning, questions that increase motivation for learners are effective. This research determines features that affect to difficulties of questions (difficulty-based features) and proposes the method for selecting questions according to the difficulty-based features for the stepwise learning. In order to manage the relations among questions, a question network is introduced in which questions are structured based on differences of each difficulty-based feature. Questions are selected by following appropriate links according the learners' answers..