Research On Educational Resource Recommendation Algorithm Based On Deep Learning And Cognitive Diagnosis

With the continuous advancement of the new generation of information technology represented by artificial intelligence,cloud computing,data mining,and mobile Internet,the rapid development of smart education and online teaching has been promoted.Smart education and online education can conveniently provide students with educational resources and help students consolidate and learn knowledge.However,with the explosive growth of the number of educational resources,students need to spend a lot of time to obtain educational resources that are truly suitable for them from the massive resources.How to fully dig the personality characteristics of student users and the hidden information in existing educational resources,improve the existing educational resource recommendation algorithm,and alleviate the problem of "information overload" and "learning lost" brought by massive educational resources has become the important research topic of modern smart education and online education.This paper analyzes the related work on students' personalized cognitive diagnosis and educational resource recommendation algorithms.Then,it investigates the improvement of the cognitive diagnosis model and the united probability matrix factorization algorithm,and implements the educational resource recommendation platform and the Office plug-ins.The details are as follows:(1)According to the cognitive diagnosis technology of educational psychology,the influential factors such as forgetting effect and number of answers on knowledge points are introduced in the calculation of key positive answer rate to improve the existing personalized student cognitive diagnosis model,combined with the students' historical records.The time sensitive deterministic inputs noisy and gate model is proposed.The experimental results show that the model is better than other cognitive diagnostic models in PMR and RMSE indicators.(2)Research on the educational resource recommendation algorithm.In order to in-depth mine the hidden information in existing educational resources and improve the accuracy of educational resources recommendation algorithm,this article integrates convolutional neural network into the united probability matrix factorization model.An educational resource recommendation algorithm based on the convolutional united probability matrix factorization model is proposed to provide students with educational resources with higher recommendation accuracy.The experimental results show that the algorithm effectively improves the accuracy;(3)According to the Spring Cloud micro-services architecture and the Office custom plug-in technology,this article researched and implemented the educational resource management and pushing method about education evaluation,which provides multiple methods for generating educational resources and docks with standard QTI format and general teaching development platform Moodle.