OPTIMISASI PARAMETER SUPPORT VECTOR MACHINE DENGAN PARTICLE SWARM OPTIMIZATION UNTUK PENINGKATAN KLASIFIKASI DIABETES

Maulana, Muhammad Reza
Sucipto, Adi
Mulyo, Harminto Mulyo

Jurnal Informatika Teknologi dan Sains (Jinteks)
Vol 6 No 4 (2024): EDISI 22

This research explores the effectiveness of using Particle Swarm Optimization (PSO) to optimize Support Vector Machine (SVM) parameters in diabetes diagnosis. Using Kaggle's "Diabetes Disease Data" dataset, this study compared the performance of SVM with default parameters and PSO-optimized SVM. Results showed small but consistent improvements in accuracy, precision, recall, and F1-score for the PSO-optimized model. Feature importance analysis identified glucose and BMI as the main predictors of diabetes. Learning curves showed both models were able to reduce overfitting as training data increased. Although the performance improvement is relatively small, this study illustrates the potential of optimization techniques in improving machine learning models for medical applications.