PENGEMBANGAN SISTEM PERAMALAN PERMINTAAN MENGGUNAKAN ALGORITMA SUPPORT VECTOR REGRESSION UNTUK OPTIMALISASI SAFETY STOCK BERBASIS WEB (STUDI KASUS: JG MOTOR SUKABUMI)

This research aims to develop a web-based system capable of forecasting motorcycle spare parts demand using the Support Vector Regression (SVR) algorithm and optimizing safety stock levels at JG Motor Sukabumi. The challenges faced in inventory management include uncertain demand fluctuations, supply delays, and the risks of overstocking or stockouts. To address these issues, the SVR model was selected due to its ability to handle non-linear and complex data, providing more accurate predictions compared to conventional methods. This study employs a descriptive quantitative approach with a quasi-experimental method to test the effectiveness of the SVR model and validate the developed web-based system.
The system is designed to offer key features, including monthly demand forecasting, safety stock calculation based on predictions, historical data visualization, and analytical reports presented through interactive graphs. The development process involves user needs analysis, the collection of two years of historical sales data, data pre-processing for normalization and standardization, SVR model training with parameter optimization, and integration of the model into a Flask-based web system. System testing was conducted using the Black Box Testing method to ensure the proper functionality of key features such as input validation, prediction processing, and stock recommendation outputs.
The results demonstrate that the SVR model achieves high prediction accuracy, evidenced by low Mean Absolute Error (MAE) values. The developed web-based system is user-friendly and enables both managers and operational staff to efficiently monitor demand and manage stock levels. Furthermore, the system has been proven to support faster and more precise strategic decision-making, helping JG Motor Sukabumi improve operational efficiency and enhance its competitiveness in the automotive market.
Keywords: Support Vector Regression, Safety Stock, Web-Based System, Black Box Testing, Demand Forecasting, Operational Efficiency.