<![CDATA[Stroke merupakan salah satu penyakit serius yang menyebabkan kematian dan kecacatan, sehingga prediksi dini terhadap risiko stroke sangat penting dalam upaya penanganan dan pencegahan. Penelitian ini mengkaji komparasi penerapan metode AdaBoost pada tiga algoritma klasifikasi, yaitu Naive Bayes, K-Nearest Neighbor (KNN), dan Decision Tree, untuk memprediksi stroke otak. Data stroke diperoleh dari dataset yang telah melalui serangkaian proses praproses, meliputi imputasi missing value, encoding variabel kategorik, dan normalisasi fitur numerik, guna meningkatkan efektivitas dan efisiensi model. Proses pelatihan dan pengujian dilakukan dengan menggunakan K-Fold Cross Validation (K=5) di platform Google Colab, dan kinerja model diukur berdasarkan metrik akurasi, presisi, recall, dan F1-Score. Hasil evaluasi menunjukkan bahwa penerapan AdaBoost secara signifikan meningkatkan performa model, terutama pada algoritma Naive Bayes dan KNN, dengan peningkatan akurasi yang mencolok; misalnya, akurasi Naive Bayes meningkat dari 82,09% menjadi 94,48% dan KNN mencapai akurasi sebesar 94,62% setelah digabungkan dengan AdaBoost. Temuan ini mengindikasikan bahwa integrasi teknik ensemble seperti AdaBoost dapat memperkuat kemampuan algoritma klasifikasi dalam mendeteksi stroke otak, sehingga berpotensi mendukung deteksi dini dan pengambilan keputusan medis yang lebih cepat serta tepat.Kata kunci: Stroke Otak, AdaBoost, Naïve Bayes, K-Nearest Neighbor, Decision Tree ABSTRACT Stroke is a serious disease that causes death and disability, so early prediction of stroke risk is very important in handling and prevention efforts. This study compares the application of the AdaBoost method to three classification algorithms, namely Naive Bayes, K-Nearest Neighbor (KNN), and Decision Tree, to predict brain stroke. Stroke data is obtained from a dataset that has gone through a series of preprocessing processes, including imputation of missing values, encoding of categorical variables, and normalization of numerical features, to improve the model's effectiveness and efficiency. The training and testing processes were conducted using K-Fold Cross Validation (K=5) on the Google Colab platform, and model performance was measured based on accuracy, precision, recall, and F1-Score metrics. The evaluation results show that the application of AdaBoost significantly improves the model performance, especially in the Naive Bayes and KNN algorithms, with a notable increase in accuracy; for example, the accuracy of Naive Bayes increased from 82.09% to 94.48% and KNN achieved an accuracy of 94.62% after incorporating AdaBoost. These findings indicate that the integration of ensemble techniques such as AdaBoost can strengthen the ability of classification algorithms to detect brain stroke, potentially supporting early detection and faster and more informed medical decision-making.Keywords: Stroke Otak, AdaBoost, Naïve Bayes, K-Nearest Neighbor, Decision Tree ]]>
Copyrights © 2025
