A Method for Predicting Egg Price Based on Multi Model Combination Optimization

With the rapid growth of the national economy and the improvement of people's living quality, eggs, as an important nutritious food, have witnessed a continuous increase in market demand. Accurately grasping the changing patterns and trends of national egg market prices is of great significance to production managers and consumers, which is conducive to promoting the stability of the laying hen industry and the social economy. This study constructs a national monthly egg market price prediction model from January 2000 to April 2024 based on a combination of methods including Seasonal-Trend decomposition using LOESS (STL), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), and Seasonal Autoregressive Integrated Moving Average (SARIMA). The results show that egg prices overall exhibit an upward trend, with seasonal and nonlinear fluctuation characteristics of “high in spring and low in autumn”. The model STL-SARIMA-SVR-LSTM constructed in this study performs best in single-step prediction of egg prices, with RMSE=0.40 and MAPE=2.68%, both of which are the minimum values.