Course knowledge graph joint embedding method introducing semantic constraint conditions

The invention relates to a course knowledge graph joint embedding method introducing semantic constraint conditions, and belongs to the field of computers. The method comprises the following steps of: 1, defining entities and relationships of a course knowledge graph to form structured data, and embedding structural information; 2, encoding directory text information in the course background information to form directory information embedding; 3, providing semantic constraint conditions under the same course category according to the brief introduction information of the course; and 4, designing a new loss function under the conditions of structure embedding, directory embedding and semantic constraint, and proposing a joint embedding method. According to the method, background information of the educational knowledge graph is fully utilized, embedding of entities and relations is more accurate, the entities can present a clustering effect in a vector space, the accuracy of entity classification tasks is impr