A Course Teacher Recommendation Method Based on Hypergraph Model
Dunhong Yao, Xiaowu Deng · 2021 International Conference on Electronic Information Technology and Smart Agriculture (ICEITSA) · 2021
Recommending suitable teachers for university courses is a personalized recommendation problem with a very practical value. However, because the recommendation task involves multiple objects and the complex cross-relationship between them, the commonly used collaborative filtering algorithm makes it difficult to recommend this data set accurately with a complex structure. The task also has the problems of cold start and data sparsity. Although the recommendation model based on an ordinary graph can use a bipartite graph to establish a simple relationship between teachers and courses, it does not reflect the complex structural relationship between courses and teachers. We propose a personalized teacher recommendation algorithm based on the hypergraph (TRH) model. The basic idea of TRH is to use the feature of hypergraph that does not lose high-dimensional information to more accurately model the complex relationship data among courses, teachers, students, supervisors, and ratings and use the manifold ranking algorithm to achieve accurate recommendations. The experimental results show that the personalized teacher recommendation algorithm based on hypergraph has higher accuracy than the user-based collaborative filtering and the recommendation algorithm based on the ordinary graph.