Construction And Application Of Curriculum Knowledge Graph In Biomedicine

With the dual background of the information age and COVID-19 new crown epidemic,various industries are undergoing intelligent changes,among which the education field is changing from offline education mode to online education(eLeaning),and the learning volume of online education platforms alone has reached 540 million people in 2020.At the same time,the main storage method of course information has been transformed from paper textbooks and teacher’s notes to graphics and videos posted on the Internet.As a big data resource in the field of education,the wisdom education platform contains a large amount of knowledge with a huge amount of hidden information,which is an important basis for promoting the development of wisdom education.How to use information technology to complete improve teaching quality,optimize teaching plans and promote educational innovation remains a hot issue in the development of smart education.The main users of the online education platform are mainly students and teachers.Students learn the knowledge points they are interested in through an independent way;teachers are responsible for the updating and optimization of the MOOC courses.In biomedical professions,the relationship between course knowledge points is more complex than the single "containment" relationship of other courses.Students often learn online courses as beginners,but the current education platform only classifies course content in the way of subjects,educational resource types and course names,ignoring the extraction and modeling of course knowledge points,so students cannot effectively and correctly construct the knowledge system of the course when learning;the knowledge points of biomedical majors are similar and common in many aspects,so teachers need to When the amount of course knowledge increases,the human-based approach tends to ignore the connection of some underlying knowledge points.Based on the above problems,the main work of this paper has the following points:Constructing a knowledge map of biomedical courses using electronic teaching materials.The order of constructing knowledge map in this paper is as follows: firstly,we extract conceptual knowledge points and inter-concept hierarchical relationships from the textbook catalog to construct the framework of course knowledge map;then we use the model based on BERT-Bi LSTM-CRF to extract knowledge point entities from a large number of textbook texts to construct the biomedical professional knowledge base;finally,we use the relationship extraction model based on Ro BERTaLSTM to extract Finally,the complex relationships between example knowledge points in the textbook are extracted using the Ro BERTa-LSTM-based relationship extraction model,and the knowledge map of the course is constructed by combining the example knowledge points and relationships with the conceptual knowledge point hierarchy.It is experimentally verified that the entity extraction model and relationship extraction model used in this paper outperform Bi Lstm+Attention and other methods.In this paper,we use Neo4 j graph database to store the extracted nodes and relations,and the total size of the knowledge graph is 1083 concept nodes and 22,296 instance nodes in total.An improved FR-GCN link prediction model based on feature fusion is designed to mine the associated knowledge points.The model integrates the global features and local features of graph data to realize the improvement of the traditional R-GCN model,and the superiority of this paper’s model is verified through experiments.The online education platform is built based on the course knowledge graph and associated knowledge point discovery method constructed in this paper.The platform is equipped with the main functions of knowledge map visualization,node search,node addition,deletion,and checking,etc.It is student group-oriented to assist in constructing the knowledge system,friendly for teaching staff to operate,and has good robustness and scalability.The system of this paper has been tested to be stable,maintainable and scalable,and can meet the usage standards.