Article Details

Abstract

In the context of smart teaching, this study aims to enhance course quality by conducting research on final grade prediction and academic risk early warning, based on data from three cohorts of students in the "Materials Physics" experimental course. For the prediction task, linear and ridge regression models effectively capture grade trends, with ridge regression demonstrating superior stability. In contrast, random forest regression and gradient boosting regression show no significant advantages under the current sample conditions. For the warning task, the logistic regression model achieves high recall but has a high false alarm rate; the support vector machine model achieves good accuracy; and the balanced random forest model performs best in minimizing missed reports. This research provides methodological and data support for teaching intervention and academic risk early warning.

Keywords

course performance; Machine Learning; smart teaching

References

1. [1] Li Yunjia, Ding Qi Research on Student Achievement Prediction Based on Machine Learning [J]. Logistics Technology, 2025, 48 (7): 178-180
2. [2] Ma Chaomin, Liu Yutong, Dou Yuhan, etc Research on predicting student academic performance based on machine learning [J]. Science and Technology Innovation and Application, 2025, 15 (25): 34-37
3. [3] Xinhua News Agency The Central Committee of the Communist Party of China and the State Council have issued the Overall Plan for Deepening the Reform of Education Evaluation in the New Era [R/OL]. (October 13, 2023) [January 3, 2026-03] https://www.gov.cn/zhengce/2020-10/13/content_5551032.htm.
4. [4] The Ministry of Education of the People's Republic of China The 14th Five Year Plan for the Development of National Education [R/OL]. (May 9, 2021) [January 7, 2022] https://www.sxie.edu.cn/bkpgw/info/1032/1815.htm.
5. [5] Sun Xin, Li Jian, Fu Zhiyu Using game log file to predict students' reasoning ability and math grades - Application of machine learning [J]. Journal of Psychology, 2018, 50 (7): 761-770
6. [6] Suarez E G，Toro R M. Teaching Performance Evaluation Model: Preparation for Student Learning Within the Framework for Teacher Good Performance [J].Journal of Educational Psychology-Propositosy Representaciones，2018，6（2）：431-452.
7. [7] Qiu Danping Research on the Application of BP Neural Network in Teaching Quality Evaluation System [J]. Modern Computer, 2020 (33): 44-46
8. [8] Yang Shilu Research on the Application of Machine Learning in Teaching Evaluation and Grade Prediction of College Physics [D]. Chengdu: University of Electronic Science and Technology of China, 2025
9. [9] Hasanah H，Farida A，Yoga P P. Implementation of Simple Linear Regression for Predicting of Students' Academic Performance in Mathematics [J].Jornal Pendidikan Matematika (Kudus)，2022：39-52.
10. [10] Homaida A，Pekalp M H，Ebegil M. Efficient k Value Computation for Enhanced Fuzzy Ridge Regression [J/OL].Journal of Computational and Applied Mathematics，2026，472：116813.https://doi.org/10.1016/j.cam.2025.116813.
11. [11] Yang Wenwei, Lu Junxiang, Bao Ruixue Tracking of SSE 50 Index Based on Ensemble Learning Model [J]. Journal of Yangzhou University (Natural Science Edition), 2025, 28 (6): 71-78
12. [12] Khaire U M，Dhanalakshmi R. Effects of Random Forest Parameters in the Selection of Biomarkers [J].The Computer Journal，2021，64（12）：1840-1847.
13. [13] Luo C Y，Chen X B，Xu J C，et al. Research on Privacy Protection of Multi-source Data Based on Improved GBDT Federated Ensemble Method with Different Metrics [J/OL].Physical Communication，2021（2）：101347.https://doi.org/10.1016/j.phycom.2021.101347.
14. [14] Xie Zhao Research and Application of Online Score Prediction Method Based on Deep Learning [D]. Guangzhou: Guangdong University of Foreign Studies, 2025
15. [15] Hemdanou A L，Sefian M L，Achtoun Y，et al. Comparative Analysis of Feature Selection and Extraction Methods for Student Performance Prediction Across Different Machine Learning Models [J/OL].Computers and Education: Artificial Intelligence，2024，7：10030.1https://doi.org/10.1016/j.caeai.2024.100301.
16. [16] Newaz A，Mohosheu M S，Noman M A A，et al. IBRF: Improved Balanced Random Forest Classifier [C]//2024 35th Conference of Open Innovations Association (FRUCT).2024：501-508.

Funding

National Natural Science Foundation of China (52472176); Undergraduate Teaching Reform Project of China University of Petroleum (East China) (CM2024086, CM2024020); Graduate Education and Teaching Reform Project of China University of Petroleum (East China) (YJG2024024)

About the authors

Li Xiaofang (1990-), female, Han ethnicity, from Qingdao, Shandong, associate professor, doctoral student, research interests include material interfaces and surfaces and enhanced oil recovery.