Design of Intelligent Diagnosis and Intervention Model of College Students’ Civic Education for New Media

Abstract

With the continuous promotion of education informatization and the rapid development of “Internet+” classroom teaching, online teaching of Civics has become an important means of ideological and political education in colleges and universities. In this study, we propose a multi-feature-enhanced deep knowledge tracking model, which utilizes recurrent neural networks to process students’ online learning data to track the changes in students’ knowledge status and further integrates the attention mechanism to enhance the predictive performance and interpretability of the model. In the experiments of predicting students’ performance, the AUC value of the DKT model with multiple features is much higher than that of the standard DKT model, and compared with the MF-DKT model, the MFA-DKT model has the most significant improvement in ACC (5.23%), AUC (3.87%) and F1 (5.99%) on the D4 dataset, which proves that the feature-rich and attentional mechanisms can effectively improve the prediction performance of the model. Improve the model’s prediction performance. The heat map of a student’s knowledge tracking results indicates that there are smooth changes in knowledge levels, which indicates that the MFA-DKT model can be easily understood. The student had the highest level of mastery of U1 and the lowest level of mastery of U4. After the precise teaching intervention, his mastery level of U4 knowledge points increased from 0.39 to 0.62, and his mastery level of U3 knowledge points increased from 0.54 to 0.83, while his weak points were strengthened.

Keywords

• recurrent neural network
• deep knowledge tracking
• attention mechanism
• university civic education
• 00a35