Demand Prediction for Food and Beverage SMEs Using SARIMAX and Weather Data

Abstract

The SME sector in Indonesia comprises 99.99% of businesses, employing 96.9% of the workforce and contributing 60.5% to GDP and non-oil exports. Despite their importance, SMEs face challenges including limited financial access, product hygiene concerns, and fluctuating demand. Accurate demand prediction is crucial for optimizing production, inventory, and resource allocation. SARIMAX and VAR models are commonly used for demand prediction, with SARIMAX proving more effective, especially when integrating weather data. Due to there are quite few literatures about SARIMAX is used at SMEs, in this study we utilized SARIMAX and VAR models with sales and weather data (average temperature and average humidity) from January to June 2023. SARIMAX with optimum parameters optimum parameters (d=1, D=1, p=2, q=3, P=2, Q=2, s=7) outperformed optimized VAR in predicting demand for food and beverage SMEs. SARIMAX obtained AIC 1070.11, MSE 80.393, MAE 7.513, RMSE 8.966 and reduced MSE by 86.35% compared to VAR. This research highlights the significance of accurate demand prediction for SMEs, emphasizing the importance of considering external factors like weather. Understanding and predicting demand patterns are vital for SMEs to make informed decisions and optimize operations efficiently.