Personalized Student Assessment based on Learning Analytics and Recommender Systems

This paper presents a process based on learning analytics and recommender systems with the objective of analyzing student assessment in order to provide clues that can help teachers in scaffolding the students' performance. For this, a set of tests was used to evaluate students' competence in direct current circuits. The tests had multiple versions and to solve them each student had to use multiple approaches. The results indicate a better performance in calculus and simulations approaches when compared with hands-on and remote laboratories approaches. The analyses also provide support for the recommendation step allowing the configuration of a knowledge base. The process as a whole is consistent in what regards its ability to make suggestions to the students as they complete a given test and to provide teachers with information that can help them formulate strategies to positively impact students' learning.