Neural Network Sales Revolution Outperforms Exponential Smoothing

Abstract

This research addresses the challenges faced by food manufacturing companies, focusing on UD. XYZ as a case study. With fluctuating sales levels causing raw material buildup and shortages, the study proposes an improved sales forecasting method to enhance raw material control. By comparing Artificial Neural Network (ANN) and Double Exponential Smoothing Holts, the research aims to optimize inventory management and production processes. Results indicate ANN's superiority over Holts, with an accuracy rate of 0.118 compared to 11.639. The ANN model accurately forecasts sales for the upcoming twelve-month period, highlighting a decline from July 2023 to May 2024. Implementing advanced forecasting methods can mitigate raw material-related risks and enhance operational efficiency for companies like UD. XYZ. Highlight:   Enhanced sales prediction methods crucial for inventory planning. Artificial Neural Network outperforms traditional forecasting techniques. Improved forecasting mitigates raw material shortages and excesses.   Keywoard: Sales forecasting, Artificial Neural Network, Raw material control, Inventory management, Production optimization.