Key points are not available for this paper at this time.
Practicing helps students reinforce the concepts that they have learned. It is equally important that personalized feedback are given to the students to help the student to learn from their mistakes. Traditionally, marking assessments and providing personalized feedback are carried out manually by teachers. However, both of these tasks become increasingly difficult as class size becomes large. To cope with this problem, we utilize natural language processing and machine learning techniques to mark essay type assessments and to help teachers provide personalized feedback promptly. The scheme determines the correctness of students' answers according to the meaning of the answers instead of comparing the words in the answers with the ones in the specimen answer. The scheme partitions the students' answers into groups according to their semantic meanings. Teachers provide feedback to a few answers in each group, and this feedback covers the typical issues demonstrated by the answers in the group. The feedback is propagated to the other answers in the same group. The proposed schemes showed reasonable marking accuracy.
Ye et al. (Tue,) studied this question.