KLASIFIKASI RISIKO PENYAKIT JANTUNG BERBASIS DATA REKAM MEDIS MENGGUNAKAN CATBOOST
Abstract
Penyakit jantung memerlukan deteksi dini agar pasien berisiko dapat diidentifikasi lebih cepat berdasarkan data klinis yang tersedia. Penelitian ini bertujuan menerapkan algoritma CatBoost untuk klasifikasi risiko penyakit jantung menggunakan data rekam medis pasien RSUD Haji Makassar. Dataset terdiri atas 640 data pasien periode 2021–2025 dengan fitur jenis kelamin, usia, glukosa, ureum, kreatinin, SGOT, SGPT, tekanan darah sistolik, dan tekanan darah diastolik. Target klasifikasi dibentuk berdasarkan diagnosis ICD-10 menjadi dua kelas, yaitu tidak berisiko dan berisiko penyakit jantung. Tahapan penelitian meliputi pembersihan data, imputasi nilai hilang menggunakan median, pembagian data latih dan data uji, pelatihan model CatBoost, serta evaluasi menggunakan accuracy, precision, recall, F1-score, confusion matrix, AUC-ROC, dan 5-fold cross-validation. Hasil pengujian menunjukkan accuracy 96,09%, weighted F1-score 94,18%, macro F1-score 49,00%, dan AUC-ROC 0,8943. Meskipun metrik agregat terlihat tinggi, confusion matrix menunjukkan model memprediksi seluruh data uji sebagai kelas berisiko dan gagal mengenali kelas tidak berisiko. Oleh karena itu, model belum layak diarahkan pada aplikasi klinis sebelum dilakukan perbaikan melalui pembobotan kelas, penyesuaian ambang keputusan, evaluasi imputasi yang lebih adaptif, dan validasi eksternal.
References
World Health Organization, “Cardiovascular diseases (CVDs),” Fact sheet, Jul. 31, 2025. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)
G. S. Collins et al., “TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods,” BMJ, vol. 385, p. e078378, 2024, doi: 10.1136/bmj-2023-078378.
L. Prokhorenkova, G. Gusev, A. Vorobev, A. V. Dorogush, and A. Gulin, “CatBoost: unbiased boosting with categorical features,” in Advances in Neural Information Processing Systems, vol. 31, 2018, pp. 6638–6648.
World Health Organization, “ICD-10 Version: 2019: International Statistical Classification of Diseases and Related Health Problems 10th Revision,” 2019. [Online]. Available: https://icd.who.int/browse10/2019/en
Z. Qiu, Y. Qiao, W. Shi, and X. Liu, “A robust framework for enhancing cardiovascular disease risk prediction using an optimized category boosting model,” Mathematical Biosciences and Engineering, vol. 21, no. 2, pp. 2943–2969, 2024, doi: 10.3934/mbe.2024131.
K. V. Tompra, G. Papageorgiou, and C. Tjortjis, “Strategic machine learning optimization for cardiovascular disease prediction and high-risk patient identification,” Algorithms, vol. 17, no. 5, p. 178, 2024, doi: 10.3390/a17050178.
scikit-learn developers, “Cross-validation: evaluating estimator performance,” scikit-learn documentation, 2026. [Online]. Available: https://scikit-learn.org/stable/modules/cross_validation.html
scikit-learn developers, “Metrics and scoring: quantifying the quality of predictions,” scikit-learn documentation, 2026. [Online]. Available: https://scikit-learn.org/stable/modules/model_evaluation.html
P. Samosir and U. Salamah, “Comparison of the performance of XGBoost, CatBoost, and GBM algorithms in cardiovascular disease prediction,” JSAI (Journal Scientific and Applied Informatics), vol. 8, no. 1, 2025.
M. M. Ali, B. K. Paul, K. Ahmed, F. M. Bui, J. M. W. Quinn, and M. A. Moni, “Heart disease prediction using supervised machine learning algorithms: Performance analysis and comparison,” Computers in Biology and Medicine, vol. 136, p. 104672, 2021, doi: 10.1016/j.compbiomed.2021.104672.
D. Asif, M. Bibi, M. S. Arif, and A. Mukheimer, “Enhancing heart disease prediction through ensemble learning techniques with hyperparameter optimization,” Algorithms, vol. 16, no. 6, p. 308, 2023, doi: 10.3390/a16060308.
T. R. Mahesh et al., “AdaBoost ensemble methods using K-fold cross validation for survivability with the early detection of heart disease,” Computational Intelligence and Neuroscience, vol. 2022, p. 9005278, 2022, doi: 10.1155/2022/9005278.
