Research On Cognitive Diagnosis Model Based On GRU Neural Network

In the era of the rise of artificial intelligence and big data,with their continuous progress and development,people increasingly expect these emerging technologies to have a certain role in promoting the field of education,and pay attention to every learner through the use of modern information means.Teaching evaluation plays an important role in the whole learning process,which can reflect the learning effect of each learner.However,only relying on traditional educators for artificial evaluation has certain one sidedness and subjectivity,which can not meet the needs of accurate evaluation of learners' personality differences.Therefore,people hope that the application of artificial intelligence and big data means,on the basis of understanding the traditional evaluation of learners,can more clearly present the knowledge structure of learners,so as to better assist parents,educators and education experts in the diagnosis of learners.Cognitive diagnosis is one of the basic problems of "AI + education".As a technical means related to education,its original intention is to discover learners' proficiency in specific knowledge concepts.The existing methods usually mine the linear interaction of students' practice process through artificially designed functions(such as logic functions),which is not enough to capture the complex relationship between students and practice.In this paper,the importance of knowledge points and the speed of doing questions are added,and a cognitive diagnosis model based on GRU network is proposed.It combines neural network to learn complex motion interaction,so as to obtain accurate and interpretable diagnosis results.Specifically,students and exercises are projected onto the factor vector,and their interaction is simulated by GRU neural network,so as to diagnose learners' mastery of knowledge points.Based on the GRU-ISCDM model,this study trained and analyzed the data of three data sets,and obtained the accuracy of learners' prediction of different practice scores,so as to determine the learners' mastery of knowledge points and knowledge level,and generate interpretable cognitive diagnosis reports in line with learners' characteristics.At the teaching level,it helps educators to better understand the learners' knowledge and make personalized teaching plans,so as to improve the teaching effect.At the level of knowledge acceptance,students can further accurately and quickly understand their existing knowledge structure and strengthen weak knowledge points.A large number of experimental results on three datasets show that the cognitive diagnosis model based on GRU neural network has better accuracy.