Vol. 1 No. 10 (2024)

This Issue 10 presents groundbreaking research on retail sales forecasting, developing an innovative prediction framework for beer category management at Supermarket X through advanced big data analytics. The study combines an improved LSTM-based deep learning model with Attention Mechanisms to process diverse data streams including sales history, temperature variations, holiday patterns, and promotional activities, achieving remarkable accuracy rates up to 98% for monthly predictions. Real-world validation across five demonstration stores demonstrated the model's practical value, delivering substantial operational improvements including a 25% increase in inventory turnover, 35% reduction in stockout rates, and 15% decrease in operating costs. This research advances retail analytics by offering practical inventory management solutions while establishing a foundation for future developments in seasonal forecasting and extreme weather response strategies.