Optimasi Algoritma Support Vector Machine untuk Prediksi Kelulusan Tes Samapta Personel Kodim 0713 Brebes

Abstract

The Samapta Test is a component of the physical fitness assessment that is mandatory for all military personnel, including personnel of the 0713 Brebes Military District Command (Kodim), as an indicator of physical readiness to carry out duties. The high failure rate of this test encourages the need for a prediction system capable of identifying personnel at risk of failure early. This study aims to build and optimize a prediction model for passing the Samapta Test using the Support Vector Machine (SVM) algorithm. Optimization is carried out through hyperparameter tuning techniques using the Grid Search Cross-Validation method to obtain the optimal combination of kernel parameters, C values, and gamma values. The data used comes from a summary of the Samapta Test results of Kodim 0713 Brebes personnel, including attributes such as age, running, push-up, sit-up, pull-up, and swimming scores. The research stages include data collection, pre-processing, feature selection, SVM model development, parameter optimization, and model performance evaluation using accuracy, precision, recall, and F1-score metrics. The results of the study show that the optimized SVM model is able to significantly increase prediction accuracy compared to the SVM model without optimization, so that it can be used as a decision-making tool for units in designing more targeted and effective personnel physical development programs.