Perbandingan Kinerja Algoritma Gradient Boosting dan LightGBM pada Prediksi Diabetes Menggunakan SMOTE dan Hyperparameter Tuning Grid Search
DOI:
https://doi.org/10.62712/juktisi.v5i2.1175Keywords:
Gradient Boosting, LightGBM, SMOTE, Hyperparameter Tuning, Diabetes PredictionAbstract
Diabetes melitus merupakan penyakit kronis dengan prevalensi yang terus meningkat secara global sehingga diperlukan sistem deteksi dini yang akurat berbasis machine learning. Penelitian ini membandingkan kinerja algoritma Gradient Boosting dan LightGBM dalam memprediksi diabetes menggunakan dua dataset publik, yaitu Diabetes Prediction Dataset (100.000 sampel) dan Early Stage Diabetes Risk Prediction Dataset (520 sampel). Eksperimen dilakukan dalam empat skenario, yaitu baseline, SMOTE, Grid Search, serta kombinasi SMOTE dan Grid Search. Evaluasi dilakukan menggunakan metrik accuracy, precision, recall, F1-score, dan AUC-ROC dengan F1-score sebagai metrik utama. Hasil penelitian menunjukkan bahwa pada Dataset 1, Gradient Boosting menghasilkan F1-score terbaik sebesar 97,01% pada skenario Baseline dan Grid Search, sedangkan LightGBM memperoleh performa terbaik pada skenario Grid Search dengan F1-score 97,00%. Pada Dataset 2, LightGBM mencapai F1-score 99,04% dan AUC-ROC 100,00% pada skenario Baseline. Selain itu, penerapan SMOTE tidak memberikan peningkatan performa yang konsisten, sedangkan efektivitas Grid Search bergantung pada karakteristik dataset yang digunakan.
Downloads
References
[1] I. Genitsaridi et al., “11th edition of the IDF Diabetes Atlas : global , regional , and national diabetes prevalence estimates for 2024 and projections for 2050,” LANCET Diabetes Endocrinol., vol. 14, no. 2, pp. 149–156, 2024, doi: https://doi.org/10.1016/S2213-8587(25)00299-2.
[2] I. D. Federation, “IDF Diabetes Atlas, 11th Edition,” Brussels, 2025. [Online]. Available: https://diabetesatlas.org/
[3] O. Iparraguirre-Villanueva, K. Espinola-Linares, R. O. Flores Castañeda, and M. Cabanillas-Carbonell, “Application of Machine Learning Models for Early Detection and Accurate Classification of Type 2 Diabetes,” Diagnotics, vol. 13, no. 14, p. 2383, 2023, [Online]. Available: https://doi.org/10.3390/diagnostics13142383
[4] A. Davinka, S. Depari, K. D. Tania, and P. E. Sevtiyuni, “Penerapan Metode Machine Learning Dan Teknik SMOTE untuk Prediksi Diabetes,” vol. 7, pp. 436–447, 2025, doi: https://doi.org/10.30865/json.v7i2.9032.
[5] D. D. Rufo, T. G. Debelee, A. Ibenthal, and W. G. Negera, “Diagnosis of Diabetes Mellitus Using Gradient Boosting,” diagnostics, vol. 11, p. 1714, 2021.
[6] M. Kivrak, U. Avci, H. Uzun, and C. Ardic, “The Impact of the SMOTE Method on Machine Learning and Ensemble Learning Performance Results in Addressing Class Imbalance in Data Used for Predicting Total Testosterone Deficiency in Type 2 Diabetes Patients and Type 2 [ 2 ]. Type 2 diabetes constitutes,” Diagnostics, vol. 14, no. 23, p. 2634, 2024.
[7] Erlin, Y. N. Marlim, Junadhi, L. Suryati, and N. Agustina, “Deteksi Dini Penyakit Diabetes Menggunakan Machine Learning dengan Algoritma Logistic Regression,” J. Nas. Tek. Elektro dan Teknol. Inf., vol. 11, no. 2, pp. 88–96, 2022.
[8] W. Nugraha and A. Sasongko, “Hyperparameter Tuning pada Algoritma Klasifikasi dengan Grid Search Hyperparameter Tuning on Classification Algorithm with,” Sist. J. Sist. Inf., vol. 11, no. 2, pp. 391–401, 2022.
[9] E. Ramadanti, D. A. Dinathi, C. Sri, K. Aditya, and D. R. Chandranegara, “Diabetes Disease Detection Classification Using Light Gradient Boosting ( LightGBM ) With Hyperparameter Tuning,” vol. 8, no. 2, pp. 956–963, 2024.
[10] U. Laila, K. Mahboob, A. W. Khan, F. Khan, and W. Taekeun, “An Ensemble Approach to Predict Early-Stage Diabetes Risk Using Machine Learning : An Empirical Study,” pp. 1–15, 2022.
[11] H. Seto et al., “Gradient boosting decision tree becomes more reliable than logistic regression in predicting probability for diabetes with big data,” Sci. Rep., vol. 12, p. 15889, 2022, doi: https://doi.org/10.1038/s41598-022-20149-z.
[12] M. R. Maulana et al., “STUDI KOMPARATIF ALGORITMA RANDOM FOREST , GRADIENT BOOSTING , DAN LIGHTGBM PADA KLASIFIKASI RISIKO DIABETES MELITUS,” vol. 10, no. 2, pp. 2116–2122, 2026.
[13] S. M. Ganie, P. K. D. Pramanik, M. B. Malik, S. Mallik, and H. Qin, “An ensemble learning approach for diabetes prediction using boosting techniques,” Front. Genet., vol. 14, no. October, pp. 1–15, 2023, doi: https://doi.org/10.3389/fgene.2023.1252159.
[14] F. Rahman, S. Hossain, J. Tiang, and A.-A. Nahid, “Diabetes Prediction Using Feature Selection Algorithms and Boosting-Based Machine Learning Classifiers,” diagnostics, vol. 15, no. 20, pp. 1–24, 2025, doi: https://doi.org/10.3390/diagnostics15202622.
[15] S. K. Nisa, M. A. Barata, and P. E. Yuwita, “Optimization of Random Forest Algorithm with SMOTE Method to Improve the Accuracy of Early Diabetes Prediction,” Sci. J. Informatics, vol. 12, no. 3, pp. 387–396, 2025, doi: https://doi.org/10.15294/sji.v12i3.22986.
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Wawan Kurniawan, Lisnawanty Lisnawanty, Kartika Handayani
This work is licensed under a Creative Commons Attribution 4.0 International License.
