Research on prediction algorithm of college students' academic performance based on Bert-GCN multi-modal data fusion
With the advent of the era of big data, predicting the academic performance of college students has become an important research topic in the education field, and traditional methods are limited to a single data source, which is difficult to fully capture the complex learning behavior of students. T...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Systems and Soft Computing |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772941925001450 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | With the advent of the era of big data, predicting the academic performance of college students has become an important research topic in the education field, and traditional methods are limited to a single data source, which is difficult to fully capture the complex learning behavior of students. To this end, this study proposes a multimodal data fusion algorithm based on BERT-GCN, which can be used to predict the academic performance of college students more accurately and help college teaching. In this study, the BERT model was used to mine deep features from text data such as course notes and assignment feedback, and then the Graph Convolutional Network (GCN) was used to integrate students' social network information and learning behavior data to construct a multimodal fusion model. Experimental data shows that the prediction accuracy of the algorithm is 92.5 %, which is 8.3 % higher than that of the traditional single-source model, and it performs well in F1 score (90.6 %) and recall rate (91.2 %). The study also shows that the multimodal data fusion method can effectively make up for the shortcomings of a single data source, significantly enhance the model's ability to understand complex learning behaviors, and outperform the existing single data source prediction methods in terms of performance. |
|---|---|
| ISSN: | 2772-9419 |