Web-Based Prediction Of Routine And Seasonal Office Supply Demand Using Linear Regression

Abstract

Inventory management of office stationery (ATK) is crucial for maintaining operational continuity and sales performance. CV Sumber Nyata still relies on manual records and past experience to plan product demand, which often leads to overstock or stockouts. This study aims to develop a web-based prediction system for routine and seasonal ATK product needs that can assist the company in determining stock requirements more systematically based on historical sales data. The research employs a quantitative method with a predictive approach using the Linear Regression algorithm, utilizing historical sales data from March 2025 to March 2026. The system was built using PHP and MySQL, and prediction accuracy was evaluated using Mean Absolute Error (MAE) and Symmetric Mean Absolute Percentage Error (SMAPE). The results show that the system successfully generates product demand predictions based on available sales data and presents the information informatively through a web interface. Black Box and White Box testing confirm that all system functions operate as required. Furthermore, user response testing from 10 respondents yielded a satisfaction rate of 89%, categorized as very good. In conclusion, the developed system can assist companies in managing ATK inventory and support decision-making for future product demand planning.