Research On Student Grade Prediction Method Based On Deep Learning

As one of the important research branches of educational data mining,the study of student achievement prediction has received extensive attention,and scholars at home and abroad have carried out some fruitful work successively.Although traditional performance prediction method has achieved good results,but there are still some shortage,these methods mainly include two aspects: one is a problem in the traditional classroom performance prediction,prediction has certain hysteresis,and data are sparse and features a single,too early for the course of teaching and management to provide effective technical support.Second,the existing online platform course score prediction research mainly uses the log data of learners’ learning on the platform,and lacks other relevant course score information.In addition,manual feature engineering is often used in modeling,which is highly dependent on the professional knowledge and experience of engineers,which affects the accuracy of prediction to a certain extent.In response to the above problems,this article uses two different data to propose different performance prediction models to improve the accuracy of performance prediction.The specific research content is as follows:1.Aiming at the problems of traditional classroom student performance prediction data in educational data mining such as sparse data and predictive lag,the performance of the courses learned in the previous semesters is used to predict the performance of the courses to be learned in the next semester or several semesters,and a fusion is proposed.Traditional classroom performance prediction model based on self-attention mechanism and depth matrix factorization.Firstly,a self-attention mechanism is added to the model,which can quickly extract important potential characteristics of students and courses,and make the model more focused on useful information.Secondly,bilinear pooling layer is constructed in the model to improve the generalization and learning ability of the model.Compared with the traditional method,this method can predict the grades before the beginning of the course and improve the predictability of the model.Experimental results show that our method is superior to the baseline comparison method.2.As for the prediction of students’ performance in online courses,the achievement of students’ performance will be affected by many factors.Aiming at the problem that the existing deep learning methods of grade prediction fail to consider the influence of multiple features on grade prediction at the same time,an online course grade prediction model combining multiple features is proposed.Firstly,the model can automatically carry out feature engineering by using deep neural network,which reduces the intervention of human feature engineering.Secondly,the model uses the factorization machine and two kinds of neural networks to consider the influence of first-order features,second-order features and higher-order features at the same time,so as to fully learn the relationship information between each feature and achievement,which improves the prediction effect of the model compared with only using a single feature to learn.The performance of the model was evaluated on the learning analysis data set of the Open University,and the experimental results show that the performance of the proposed model is better than that of the existing model.