How Do Fluctuations in Raw Material Prices Affect the Final Product Prices? Can This Be Predicted Utilizing AI Models?

Globally, prices are soaring, and among these, agricultural product prices are experiencing significant increases. Agricultural products are essential for human survival and have a direct impact, making this a highly sensitive issue. This problem has been further exacerbated by the COVID-19 pandemic and the Ukraine-Russia war, leading to even greater price hikes. Consequently, this situation places a considerable burden not only on consumers but also on the companies that process and produce these agricultural products. In particular, South Korea, a country with a high dependency on imports, relies on foreign markets for many agricultural products. However, despite this reliance, there is a lack of research on predicting the prices of imported agricultural products. For this reason, this study implements an AI model that predicts the retail prices of final products based on fluctuations in the prices of imported agricultural products, focusing on key research questions (RQs) and the distribution process. In addition, it examines the impact of global issues such as supply chain disruptions caused by COVID-19 and the Ukraine-Russia war on price inflation. In this study, the top three imported agricultural products―corn, wheat, and soybeans―are selected as key imported agricultural items, and the prediction model is implemented focusing on the processed foods most commonly consumed in daily life that are manufactured from these products. At this time, key variables are added by taking into account the distribution and manufacturing processes involved in the processing of agricultural products. Specifically, the relationships between these variables are analyzed through correlation analysis, and a prediction model is implemented using big data analysis to select the model with the highest prediction accuracy. In this study, the VAR model showed the highest prediction accuracy among machine learning models, while the LSTM model demonstrated the highest accuracy among deep learning models. In conclusion, the academic implications of this study demonstrate that the impact of the international commodity market and global trade on the production and distribution of domestic companies can be scientifically analyzed and proposed through AI models. Additionally, the study systematically identified and explained the raw material supply disruptions caused by the COVID-19 pandemic from an academic perspective. Furthermore, the practical implications of this study for companies are significant. By utilizing this research, companies can identify how price fluctuations in imported agricultural products affect the pricing of final products. It allows them to proactively respond to changes in the prices of final products by applying the forecasting model. Additionally, companies can confirm the relationship between the prices of imported agricultural products and the retail prices of final products, gaining a clearer understanding of the costs incurred during domestic and international distribution processes. It can be used to drive innovations in the distribution structure.