Predicting Student Performance with Deep Learning
Author Information
Author(s): Tang Bo, Li Senlin, Zhao Changhua
Primary Institution: Huaihua University
Hypothesis
Can deep ensemble learning accurately predict student performance?
Conclusion
The proposed method significantly improves the accuracy of predicting student performance compared to traditional models.
Supporting Evidence
- The proposed method achieved an RMSE of 1.6562 and a MAPE of 9.7532.
- Feature selection techniques improved model performance by reducing noise.
- The ensemble approach allowed for better integration of predictions from multiple models.
Takeaway
This study shows a new way to guess how well students will do in school using smart computer programs that learn from data.
Methodology
The study used deep ensemble learning with feature ranking and optimization techniques to predict student performance based on data collected from questionnaires.
Potential Biases
Potential bias from self-reported data and the specific demographic of the sample.
Limitations
The study's findings may not be generalizable due to the specific region and limited age range of the participants.
Participant Demographics
628 students from two technical and engineering faculties in Nanjing, China, with a mean age of 23.78 years.
Statistical Information
P-Value
1.6562
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website