Analysis and Implementations of Support Vector Machine for Predicting the Number of Ship Passengers at PT Pelni Medan Branch

Abstract

This study examines the application of the Support Vector Machine (SVM) algorithm to predict the number of ship passengers at PT Pelni Medan Branch. The main problem faced by the company is the difficulty in forecasting passenger numbers due to fluctuations influenced by seasonal factors, economic conditions, and special events. Until now, the company has not had a reliable data-based prediction system, so operational decisions are still made manually, which often leads to capacity imbalances. The SVM algorithm was chosen for its ability to perform both linear and non-linear predictions by utilizing a hyperplane as a class separator. Research data were obtained from historical passenger records, which were processed through preprocessing, splitting into training and testing sets, and SVM model training. The implementation was realized through a web-based application using PHP and MySQL to make prediction results easily accessible to the company. The findings show that the SVM model can classify passenger numbers into three categories: low, medium, and high. The system was tested using the Blackbox Testing method and declared valid for all tested functions. Thus, this research contributes to supporting operational decision-making and planning at PT Pelni Medan Branch, although it is still limited by the number of variables used.