Deep Learning Algorithms for Predicting Student Course Performance and Improving Teacher Guidance

This article is based on the method of deep learning. Based on previous research on the prediction of MOOCs course performance, multi-dimensional analysis of experimental data is conducted to analyze the correlation between experimental data and courses, the performance of courses, the trend of course performance, and the reasons for course performance. Data visualization and fitting methods are used to determine the feature extraction items in course performance prediction, And modeling based on a deep neural network model. This article proposes three different training methods, including direct training method, sub course training method based on course grouping after course association analysis, and time series prediction method based on the time attribute characteristics of experimental datasets. Finally, the feasibility of course performance prediction was verified through experiments.