Research on Course Recommendation Algorithm Based on Knowledge Graph

In recent years, with the development of information technology, information resources on the Internet have increased explosively. A variety of online education platforms have also appeared in front of learners, providing learners with sufficient reference materials and learning materials, bringing a lot of convenience. Everything has two sides. The development of information technology not only brings convenience to learners, but also brings a lot of problems to learners. The number of complex learning resources on the Internet leads to too many choices of learners and excessive information, which increases the dropout rate of learners and reduces the passing rate of courses. Therefore, many education platforms use recommendation systems to recommend popular courses, but popular courses are not necessarily suitable for learners, so personalized recommendation systems have emerged. But traditional recommendation system has some problems, sparse data and cold boot, et al. Therefore, it is necessary to use auxiliary information to assist the recommendation system to make course recommendation to alleviate these problems. Therefore, this paper proposes a new model. The model consists of three parts: the left part is the prediction part of the recommender system, the middle part models the left and right two parts using the cross-compression unit, and the right is a good item representing the vector through GCN. Through CTR, we verify the experimental results of AUC, ACC, HR and NDCG, and compare them with other algorithms.