INTEGRASI ALGORITMA K-NEAREST NEIGHBORS DAN DECISION TREE UNTUK MEMPREDIKSI HIPERTENSI

Hypertension is a prevalent health condition and a major risk factor for cardiovascular diseases. Early detection and management are essential to prevent complications. This study aims to optimize the accuracy and stability of hypertension risk prediction by applying a stacked ensemble technique that combines multiple base classifiers—K-Nearest Neighbors (KNN) and Decision Tree (DT)—with Logistic Regression as the meta-learner. The dataset used was imbalanced, thus requiring class balancing with the Synthetic Minority Over-sampling Technique (SMOTE), along with data preprocessing and scaling. The study applies a quantitative approach to train and evaluate models using Python. Results demonstrate that the stacked ensemble model achieves superior performance compared to individual classifiers, with a maximum accuracy of 74.52%. These findings indicate that the combination of different classifiers through ensemble stacking enhances the reliability and predictive capability of hypertension detection models. The approach offers potential value for improving early diagnosis and supporting clinical decision-making.