Optimasi Metode Support Vector Machine (SVM) Mengunakan Particle Swarm Optimization pada Permasalahan Klasifikasi Diabetes

Abstract

Diabetes mellitus is a chronic disease that requires accurate early detection. This study presents a diabetes classification system by integrating Support Vector Machine (SVM) with Particle Swarm Optimization (PSO) to automatically optimize model parameters. The dataset used was obtained from Kaggle, consisting of 100,000 entries and nine medical attributes. Data preprocessing included cleaning, encoding, Min-Max normalization, and undersampling to balance class distribution. Model performance was evaluated using 5-Fold Cross Validation. The results showed that the SVM- PSO achieved an average accuracy of 83.60% which is higher than the conventional SVM with 83.39% accuracy. These findings demonstrate that PSO effectively enhances the classification performance of SVM and is recommended for machine learning-based medical diagnosis, especially in diabetes prediction.