육계의 생애주기 및 AI를 고려한 수급 및 가격 예측모형 개발을 위한 연구

Abstract

This study analyzes the supply-demand and price prediction model considering the life cycle (breeding stage) of broilers. For this purpose, the supply and demand and price were predicted by considering the life cycle of broilers from the Parents Stock to the slaughter stage using Vector Error Correction Model and the Gated Recurrent Unit, based on the data for the last 10 years. Factors that did not affect the dependent variable were excluded through multiple regression analysis during the prediction process. As a result of the supply and demand forecasting, it was found that the forecast results through Vector Error Correction Model showed a small deviation compared to Gated Recurrent Unit. However, in the case of Gated Recurrent Unit, However, in the case of Gated Recurrent Unit, it was found that the accuracy of the prediction values and the similarity of the patterns improved as the number of learning times increased. On the other hand, in both models, the accuracy of the price prediction result was found to be quite low. For more accurate prediction, it is necessary to increase the size of the data through shortening of the data generation cycle and data segmentation for each broiler life cycle.