Komparasi Algoritma Machine Learning untuk Prediksi Risiko Burnout Akademis Akibat Penggunaan Generative AI
DOI:
https://doi.org/10.62712/juktisi.v5i2.1366Keywords:
Burnout Akademis, Generative AI, Machine Learning, Random Forest, XGBoost, Feature ImportanceAbstract
Perkembangan Generative AI (GenAI) dalam pendidikan tinggi memunculkan fenomena baru berupa peningkatan risiko burnout akademis mahasiswa yang tidak lagi semata dipicu oleh beban belajar konvensional, melainkan oleh pola interaksi dengan kecerdasan buatan. Penelitian ini bertujuan membandingkan kinerja tiga algoritma Machine Learning, yaitu Logistic Regression, Random Forest Classifier, dan XGBoost, dalam mengklasifikasikan tingkat risiko burnout akademis (High, Medium, Low) berdasarkan data profil 50.000 mahasiswa dengan 16 atribut, termasuk durasi penggunaan GenAI mingguan (Weekly_GenAI_Hours) dan tingkat ketergantungan psikologis terhadap AI (Perceived_AI_Dependency). Metode penelitian meliputi eksplorasi data, pra-pemrosesan (encoding, normalisasi, pembagian data latih-uji 80:20), pelatihan model baseline, optimasi hyperparameter menggunakan GridSearchCV, serta evaluasi menggunakan metrik akurasi, presisi, recall, F1-score, confusion matrix, dan feature importance. Hasil penelitian menunjukkan akurasi model berkisar antara 51,96% hingga 53,83%, dengan Logistic Regression baseline mencatatkan akurasi tertinggi (53,83%), diikuti XGBoost setelah tuning (53,62%) dan Random Forest setelah tuning (53,27%). Hyperparameter tuning terbukti meningkatkan akurasi model berbasis pohon secara signifikan, namun tidak pada model linear. Ekstraksi feature importance dari Random Forest dan XGBoost secara konsisten menempatkan Weekly_GenAI_Hours dan Perceived_AI_Dependency sebagai dua prediktor terkuat, jauh melampaui variabel akademik tradisional seperti IPK dan jam belajar. Analisis confusion matrix mengungkap tumpang tindih klasifikasi terutama pada kelas Medium. Temuan ini memberikan bukti empiris bahwa intensitas dan ketergantungan penggunaan GenAI merupakan determinan utama burnout akademis, sehingga dapat menjadi dasar objektif bagi kebijakan pembatasan penggunaan AI yang sehat di lingkungan pendidikan.
Downloads
References
[1] E. Yeskuatov, L. K. Foo, and S.-L. Chua, "Detecting Burnout Among Undergraduate Computing Students with Supervised Machine Learning," Healthcare, vol. 13, no. 23, p. 3182, 2025, doi: 10.3390/healthcare13233182.
[2] R. Pereira et al., "Burnout Risk Profiles in Psychology Students: An Exploratory Study with Machine Learning," Behavioral Sciences, vol. 15, no. 4, p. 505, 2025, doi: 10.3390/bs15040505.
[3] G. Feher, K. Kapus, A. Tibold, Z. Banko, G. Berke, B. Gacs, I. Varadi, R. Nyulas, and A. Matuz, "Mental issues, internet addiction and quality of life predict burnout among Hungarian teachers: a machine learning analysis," BMC Public Health, vol. 24, p. 2322, 2024, doi: 10.1186/s12889-024-19797-9.
[4] P. Sharma et al., "Hybrid Model for Early Detection of Student Burnout," IEEE Transactions on Learning Technologies, 2024.
[5] N. A. Kuadey, C. Ankora, F. Tahiru, L. Bensah, C. C. M. Agbesi, and S. O. Bolatimi, "Using machine learning algorithms to examine the impact of technostress creators on student learning burnout and perceived academic performance," International Journal of Information Technology, vol. 16, no. 4, pp. 2467-2482, 2024, doi: 10.1007/s41870-023-01655-3.
[6] H. Nellore, "Data Analytics and Machine Learning Approaches for Predicting Academic Stress and Enhancing Student Wellbeing," International Journal of Advance Research and Innovation, vol. 13, no. 4, pp. 23-29, 2025, doi: 10.69996/ijari.2025021.
[7] R. P. Tapio, "Predicting Burnout in College Students: A Machine Learning Approach Using Decision Tree and Psychometric Data," Asian Journal of Probability and Statistics, vol. 27, no. 5, pp. 50-60, 2025, doi: 10.9734/ajpas/2025/v27i5754.
[8] M. Jukiewicz, "How generative artificial intelligence transforms teaching and influences student wellbeing in future education," Frontiers in Education, vol. 10, 2025, doi: 10.3389/feduc.2025.1594572.
[9] M. Besalti, "Harnessing Self-Control and AI: Understanding ChatGPT's Impact on Academic Wellbeing," Behavioral Sciences, vol. 15, no. 9, p. 1181, 2025, doi: 10.3390/bs15091181.
[10] T. Chen and C. Guestrin, "XGBoost: A Scalable Tree Boosting System," in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, San Francisco, CA, USA, 2016, pp. 785-794, doi: 10.1145/2939672.2939785.
[11] L. Breiman, "Random Forests," Machine Learning, vol. 45, no. 1, pp. 5-32, 2001, doi: 10.1023/A:1010933404324.
[12] N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, "SMOTE: Synthetic Minority Over-sampling Technique," Journal of Artificial Intelligence Research, vol. 16, pp. 321-357, 2002, doi: 10.1613/jair.953.
[13] J. Tian and R. Zhang, "Learners' AI dependence and critical thinking: The psychological mechanism of fatigue and the social buffering role of AI literacy," Acta Psychologica, vol. 260, p. 105725, 2025, doi: 10.1016/j.actpsy.2025.105725.
[14] F. Pedregosa et al., "Scikit-learn: Machine Learning in Python," Journal of Machine Learning Research, vol. 12, pp. 2825-2830, 2011.
[15] D. M. W. Powers, "Evaluation: From Precision, Recall and F-Measure to ROC, Informedness, Markedness and Correlation," Journal of Machine Learning Technologies, vol. 2, no. 1, pp. 37-63, 2011.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Ophelius Savero Rinyuakng Rolin, Wahyu Nugraha, Rabiatus Sa'adah
This work is licensed under a Creative Commons Attribution 4.0 International License.
