The rapid and vigorous development of Internet technology,especially the popularity of mobile terminals,provides users with a large amount of information resources.Users are gradually used to doing many things on the Internet,such as buying goods,watching film and television works,listening to popular music and viewing the latest information.However,this large amount of resources will make users lose themselves in the dazzling information.Therefore,if users want to find the content they like or are interested in in the large amount of information,they need to spend more time than before.The fundamental purpose of recommendation system is to build the relationship between users and items according to the historical behavior data of users.Research in academia and industry has proved that the recommendation system can more effectively help users retrieve content similar to their interests and hobbies in a large amount of information,so as to improve the effective use of big data.However,affected by the small amount of model structure and feature information,the traditional recommendation algorithms often can not achieve the ideal effect.Due to the comprehensive auxiliary data of knowledge graph,it has attracted more and more attention.It usually represents the semantic information of entities and relationships between entities in the form of triples,which can accurately express the semantic information of user interaction with the item.In recent years,the types of elective curriculums for college students have become more and more rich and diverse,which makes it difficult for students to choose curriculums they want.This is because the information overload leads students to blindly choose curriculums without knowing whether the curriculum content is in line with their own development and consistent with their own interests.Therefore,in this paper,we propose a model of student curriculums selection recommendation system based on knowledge graph and graph convolutional network,which aims to provide students with personalized curriculums selection recommendation.The main research of this paper includes:(1)This paper proposes a model of student curriculum selection recommendation system based on graph convolutional network.The purpose of this study is to input the knowledge graph of student curriculum selection into the graph convolutional network architecture,use the graph convolutional network to propagate and embed from the neighbor of the node to update the new representation of the node,then aggregate the representation of users and items,and output the predicted matching score,So as to realize students’ Personalized Curriculum Selection recommendation.(2)Based on the model of student course selection recommendation system based on graph convolutional network,we introduce the double attention mechanism,that is,the combination of double attention mechanism and graph convolution network,which not only considers the characteristic relationship between the current node and its neighbor nodes,but also takes into account the characteristic relationship between neighbor nodes and the current node,and makes full use of the relationship characteristics between the current node and the neighbor nodes,This enhances the accuracy,personalization and interpretability of students’ course selection recommendation. |