Student Profile Clustering Based Personalized Exercise Recommendation: Taking Data Structures Course as an Example

Zhaowei Zhang, Lin Liu, Hui Liu, Zhenyu Wu

Abstract: In data structures course, the numerous knowledge points pose a great challenge for all students. Also, the students' different individual conditions cause the different scores over different knowledge points. To increase the student's scores, the personalized exercise recommendation is useful. However, the traditional personalized exercise recommendation brings about a huge waste of time and decreases the teaching efficiency for teachers. Introducing the clustering idea, we propose one profile clustering based personalized exercise recommendation (PCPER). Based on all knowledge points, this method first conducts all students' profiles, then executes the clustering over all profiles and finally recommends the personalized exercise for all clusters. The students falling into one cluster receive the same exercise for one knowledge point. The preliminary results show that the proposed PCPER method not only increases the student's scores over all knowledge points but also saves the teacher's time.