Sistem Prediksi Penjualan Toko Kelontong Berbasis Website Dengan Metode Holt-Winters Exponential Smoothing

Abstract

The development of information technology provides opportunities for business owners to improve the effectiveness of business management, including inventory management. Grocery stores often face problems related to overstock and stock shortages due to the absence of a sales forecasting system capable of predicting future product demand. This study aims to develop a web-based sales forecasting system for grocery stores using the Holt-Winters Exponential Smoothing method and to determine the accuracy level of the resulting forecasting model. The system was developed using PHP and Python programming languages with MySQL as the database management system. PHP was used for interface and database management, while Python was used for the forecasting computation process. The Holt-Winters Exponential Smoothing method was applied because it is capable of handling time series data with trend and seasonal patterns. The alpha (α), beta (β), and gamma (γ) parameters were automatically optimized using the Limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS-B) optimization algorithm. Model accuracy was evaluated using Mean Absolute Percentage Error (MAPE), while system functionality testing was conducted using the Black Box Testing method. The results of this study indicate that the system successfully manages product and sales data, as well as displays sales forecasting results in the form of tables and interactive charts. Based on Black Box Testing, all system features functioned according to the specified functional requirements. Furthermore, the accuracy evaluation showed that the Holt-Winters Exponential Smoothing method was able to produce MAPE values categorized as good to very good for several tested products. Therefore, the developed system can assist store owners in forecasting sales and support decision-making related to inventory management