Learning path recommendation based on Transformer reordering
Yunxiang Liu, Yuanyuan Zhang, Guoqing Zhang · 2020 5th International Conference on Information Science, Computer Technology and Transportation (ISCTT) · 2020
In order to solve the problem that course learning resources cannot be accurately recommended, this paper proposes learning path recommendation based on transformer reordering. This study constructs the map of knowledge points in the curriculum in advance. When the learners request to learn new knowledge points, the DINA cognitive model is used to analyze the test results, and the blind area knowledge points are obtained and the map of blind area knowledge points is constructed. Then, RankNet algorithm based on neural network is used to realize the first sorting of knowledge points in blind area, and then the personalized characteristics of learners are introduced into transformer algorithm to reorder. Finally, the recommender sequence is combined with the map of knowledge points in the blind area, and the recommended sequence is generated by topological sorting method. In this study, knowledge model and transformer algorithm are used to make up for the deficiency of Curriculum Resource Recommendation problem in previous studies. Taking high school chemistry knowledge points recommendation as an example, the experimental results show that the satisfaction of this method reaches 82%. This recommendation method promotes the research process of personalized learning resources to a certain extent.