Exercises Recommendation in Adaptive Learning System
Zuoxi Jin, Kun Ma, Kun Liu, Ke Ji · 2019
The adaptive learning system develops gradually, but most attention is paid to the construction of student model and domain model. In this paper, a recommendation algorithm based on students' current knowledge level is proposed to match suitable exercises and avoid homogenization of learning content for all students, for the purpose of achieving so-called "adaptative". It is worth noting that the learning system recommendation is different from the general recommendation. Not only the method, the evaluation standard of recommendation result is also different. We should not simply recommend to students the exercises they must or must not mastered, but recommend to them the learning resources they should have within the range of their abilities according to the theory of proximal development zone. We also use the bayesian knowledge tracing model to judge students' mastery of knowledge as the evaluation standard of this algorithm.