Cognitive Diagnosis And Recommendation System Based On Knowledge Graph Research And Application Realization

In recent years,with the development of artificial intelligence technology,more and more researchers apply artificial intelligence technology to the field of education.With its unique technical advantages,artificial intelligence has set off a wave of upsurge in the education industry.Due to the increasing update of the Internet industry,online learning and online courses have become an indispensable part of most students’ learning and life,which has promoted the development of a new generation of intelligent education online platform.Different from the traditional MOOC platform,which focuses on providing learning resources,the new generation of intelligent online education platform provides personalized education services for different individuals,and the Knowledge Graph(KG)is the core component of the new generation of intelligent online education platform.In addition,the development of computer-assisted learning systems has promoted the study of cognitive diagnosis,enabling cognitive diagnostic techniques to predict the performance of students in course work over time.Based on the existing education auxiliary system,this paper studies the methods of integrating knowledge graph,cognitive diagnosis technology and personalized recommendation algorithm.The specific work is as follows:1.A cognitive diagnosis algorithm model based on knowledge graph is constructed.This paper firstly obtains knowledge points through web crawler,then constructs subject knowledge map based on junior and senior high school mathematics,and finally judges students’ knowledge state through cognitive diagnosis model.In the part of cognitive diagnosis model,the Knowledge Graph Item Response Theory(KGIRT)model proposed in this paper completes the cognitive assessment task of predicting students’ performance,and a large number of experiments on large-scale real data sets prove the effectiveness and explanatory ability of the system.2.A recommendation model based on the cryptic meaning algorithm is built.In this paper,a Deep Latent Factor Model with Hierarchical Similarity(DLFM-HS)is designed to solve the problem of rating prediction.At the same time,in order to better measure the similarity between the user and the project,this paper designs a hierarchical similarity measure to replace the inner product which is widely used now.In addition,in order to reduce the sensitivity of the model to data sparsity,this paper adopts deep neural network to learn user preferences and project Outlines from project description.Experimental results on five real data sets show that the DLFM-HS model has higher accuracy than the traditional model.3.On the basis of the application of knowledge graph and cognitive diagnosis model,a We Chat applet-Deep Assistant AI education system was developed,which realized the transformation from theory to application.After completing the cognitive detection,the system uses the knowledge graph to obtain the relevant knowledge points and make personalized recommendation of learning resources according to the students’ mastery of the current knowledge points.In addition,in order to facilitate the management of teaching content and teachers,this paper develops a content management platform,which realizes the functions of knowledge point management,video management,question bank management and teacher management.