教育与研究

基于可解释机器学习的青少年厌学风险识别与关键预测因素分析

崔 乐康

浙大宁波理工学院

刘 璇

浙大宁波理工学院

阮 晨阳

浙大宁波理工学院

周 晚

浙大宁波理工学院

王 欢

浙大宁波理工学院

左 万里

宁波大学

DOI: https://doi.org/10.59429/jyyj.v8i14.15259

Keywords: 厌学；机器学习；PISA 2022；风险预测；SHAP

---

Abstract

厌学风险受到个体、家庭与学校等多层因素共同影响，传统线性方法难以充分识别复杂非线性关系。本研究 基于 PISA 2022 数据，从学校环境、家庭背景和个体特征三个维度构建 13 项预测指标，采用决策树、逻辑回归、K 近邻、支持向量机和 CatBoost 进行分类，并以 SHAP 解释模型。结果显示，五种模型 AUC 均超过 0.90 ；KNN 在 AUC 和 F1-score 方面表现最佳。SHAP 分析表明，留级经历、毅力与坚持性、合作能力、换位思考能力和父母情感 支持是 CatBoost 预测的重要特征。研究为青少年厌学风险的早期识别与学生支持提供数据驱动参考。

---

References

[1] FREDRICKS J A, BLUMENFELD P C, PARIS A H. School engagement: Potential of the concept, state of the evidence[J]. Review of Educational Research, 2004, 74(1): 59-109.

[2] FINN J D, ZIMMER K S. Student engagement:What is it? Why does it matter?[M]//CHRISTENSON S L,RESCHLY A L, WYLIE C, eds. Handbook of Research on Student Engagement. Boston: Springer, 2012: 97-131.

[3] PUTWAIN D W, WOOD P, PEKRUN R.Achievement emotions and academic achievement: Reciprocal relations and the moderating influence of academic buoyancy[J].Journal of Educational Psychology, 2022, 114(1): 108-126.

[4] BRONFENBRENNER U. The Ecology of Human Development: Experiments by Nature and Design[M].Cambridge, MA: Harvard University Press, 1979.

[5] ECCLES J S, ROESER R W. Schools as developmental contexts during adolescence[J]. Journal of Research on Adolescence, 2011, 21(1): 225-241.

[6] BAKER R S, INVENTADO P S. Educational data mining and learning analytics[M]//LARUSSON J A, WHITE B, eds. Learning Analytics: From Research to Practice. New York: Springer, 2014: 61-75.

[7] BREIMAN L. Random forests[J]. Machine Learning,2001, 45: 5-32.

[8] FRIEDMAN J H. Greedy function approximation:A gradient boosting machine[J]. The Annals of Statistics, 2001,29(5): 1189-1232.

[9] CHEN T, GUESTRIN C. XGBoost: A scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2016: 785-794.

[10] LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions[C]//Advances in Neural Information Processing Systems 30. 2017: 4765-4774.

[11] WENTZEL K R. Teacher-student relationships[M]//WENTZEL K R, MIELE D B, eds. Handbook of Motivation at School. 2nd ed. New York: Routledge, 2016.

[12] ALLEN K A, KERN M L, VELLA-BRODRICK D, et al. What schools need to know about fostering school belonging: A meta-analysis[J]. Educational Psychology Review,2018, 30(1): 1-34.

[13] OECD. PISA 2022 Assessment and Analytical Framework[R]. Paris: OECD Publishing, 2023.