Construction Of Student Model Based On Cognitive Diagnosis

With the continuous deepening of education informatization and the gradual introduction of the national double reduction policy in recent years,how to use a large amount of collected online and offline education data to mine and analyze the shortcomings and potential of each student,give Accurate and interpretable personalized diagnostic results,and providing each student with customized learning content and methods based on the results,has become the current research focus of personalized intelligent education.However,there are two difficulties in establishing a targeted student personalization model and test item recommendation system: the first is how to ensure the interpretability of the diagnosis model on the basis of improving the accuracy of students’ cognitive diagnosis results;the second is how to In the actual teaching process,combined with the students’ diagnosis results,personalized test questions are recommended.In this paper,a systematic and in-depth study on these two difficulties is carried out,and the main research results are as follows;1.A student cognitive diagnosis model combined with graph network is established.By combining neural network and traditional cognitive diagnosis methods,the accuracy of diagnosis results is improved while ensuring the interpretability of results;2.A personalized test item recommendation model based on cognitive diagnosis is established.In the traditional test item recommendation system,this method combines the cognitive diagnosis results of students in the previous step to achieve personalized and hierarchical test item recommendation;3.The above two models are tested on real data sets for many times.The experimental results prove the effectiveness of the model for students’ cognitive diagnosis and personalized test question recommendation.Finally,the application process of the two models in actual teaching is shown.