PERS: A Personalized Recommender System for Student-Generated Questions in Programming Courses

Abstract

This study introduces PERS (Personalized Exercise Recommender System), integrating Student-Generated Questions (SGQ) with a recommender system for programming courses. Implemented in a 10-week introductory programming course with 395 undergraduates (200 using PERS, 195 as control), PERS uses collaborative filtering to suggest personalized exercises. This study addresses the impact of PERS on student engagement and learning outcomes, the effectiveness of its recommender system, and students' perceptions of the system. Results show significant improvements in exercise completion rates, assignment scores, and final exam performance for PERS users compared to the control group. The system demonstrated high recommendation relevance and user satisfaction, suggesting its potential to enhance personalized learning in programming education.