Application of Ensemble Method to Predict Individual Pork Prices Using Multi-Source Information

Kyungchang Jeong, Eunyoung Ko, Gyuchang Jo, Hongseok Oh, Ji-Hoon Jeong, Yuan H. Brad Kim, Jungseok Choi, Euijong Lee

Abstract: The role of pork in the food industry’s supply chain is crucial, and the price of pork has a significant impact on both consumers’ quality of life and the swine industry. Previous research has attempted to improve the accuracy of pork price trend predictions by incorporating various external factors and enhancing artificial neural networks. In contrast, this study aims to predict the auction price of individual pork, enabling real-time price predicting at the modern slaughterhouse. To achieve this, multi-source data were gathered, including the characteristics of individual pork and external factors influencing pork prices. This study proposes a stacking ensemble method that combines multiple machine learning and deep learning models, leveraging their diverse strengths to enhance overall performance and improve generalization to unseen data. The experimental results demonstrate that the proposed method not only achieves the lowest mean absolute percentage error of 3.262, but also accurately predicts individual pork prices and outperforms standalone machine learning and deep learning models across various scenarios.

External IDs:dblp:journals/access/JeongKJOJKCL25