KLASIFIKASI FAKTOR RISIKO PENYAKIT JANTUNG MENGGUNAKAN MACHINE LEARNING

CLASSIFICATION OF HEART DISEASE RISK FACTORS USING MACHINE LEARNING

Abstract

Penyakit jantung merupakan salah satu penyebab utama kematian di seluruh dunia, termasuk di Indonesia. Deteksi dini terhadap risiko penyakit jantung sangat penting untuk mencegah komplikasi serius namun tantangan utama yang dihadapi adalah deteksi dini penyakit ini sering terlambat karena proses klasifikasinya belum cukup akurat. Penelitian ini bertujuan membangun model klasifikasi risiko penyakit jantung dengan pendekatan machine learning. Dataset yang digunakan berasal dari Kaggle, terdiri dari 920 data pasien dengan 13 fitur medis yang relevan. Proses penelitian meliputi pra-pemrosesan data, seleksi fitur menggunakan metode Recursive Feature Elimination (RFE), pelatihan model dengan enam algoritma (Random Forest, Support Vector Machine, K-Nearest Neighbors, Naive Bayes, Decision Tree, dan Logistic Regression), serta evaluasi kinerja model menggunakan metrik akurasi, precision, recall, f1-score, ROC-AUC, dan analisis tingkat kepentingan fitur. Hasil penelitian menunjukkan bahwa algoritma Support Vector Machine (SVM) memberikan performa terbaik dengan akurasi sebesar 83,91% dan AUC sebesar 0,92, diikuti oleh Random Forest dengan akurasi 82,61%. Fitur yang paling berkontribusi terhadap hasil klasifikasi adalah tipe nyeri dada (cp), jumlah pembuluh darah utama (ca), dan jenis talasemia (thal). Temuan ini menunjukkan bahwa penerapan machine learning dengan seleksi fitur yang tepat dapat meningkatkan akurasi klasifikasi risiko penyakit jantung dan berpotensi digunakan sebagai sistem pendukung keputusan dalam dunia medis.

Heart disease is one of the leading causes of death worldwide, including in Indonesia. Early detection of heart disease risk is very important to prevent serious complications but the main challenge faced is that early detection of this disease is often late because the classification process is not accurate enough. This study aims to build a heart disease risk classification model with a machine learning approach. The dataset used comes from Kaggle, consisting of 920 patient data with 13 relevant medical features. The research process includes data pre-processing, feature selection using the Recursive Feature Elimination (RFE) method, model training with six algorithms (Random Forest, Support Vector Machine, K-Nearest Neighbors, Naive Bayes, Decision Tree, and Logistic Regression), and model performance evaluation using accuracy, precision, recall, f1-score, ROC-AUC metrics, and feature importance analysis. The results showed that the Support Vector Machine (SVM) algorithm provided the best performance with an accuracy of 83.91% and AUC of 0.92, followed by Random Forest with an accuracy of 82.61%. The features that contributed most to the classification results were chest pain type (cp), number of major blood vessels (ca), and thalassemia type (thal). These findings suggest that the application of machine learning with appropriate feature selection can improve the accuracy of heart disease risk classification and has the potential to be used as a decision support system in the medical world.