Research On Egg Price Prediction Method Combining Intervention Analysis And Multiple Factors

As residents’ consumption level in China improves and lifestyles change,the demand for high-protein products has grown rapidly,with eggs and their related products being one of the most important protein sources for Chinese residents.The abnormal fluctuations in egg prices caused by external factors are directly related to the stability of the egg market,related industries,and the vital interests of the general public.As a link in many industry chains,price fluctuations of upstream and downstream related products and unforeseeable emergencies can significantly impact the egg market.Based on existing price prediction methods,this thesis combines an intervention analysis model.It considers multiple influencing factors to build a composite model for predicting future egg prices and designs and implements an egg price fluctuation trend prediction system.The main work is as follows:Single model prediction and the construction of the egg price index system: First,a SARIMA model is established to predict the linear trend in the egg price sequence,obtaining the error sequence.Second,an analysis of the causes of egg price fluctuations is conducted,and the main influencing factors are selected.The gray relational degree analysis determines their correlation with egg prices,providing a basis for establishing a multi-feature prediction model.Building a composite model capable of effectively fitting errors and conducting experimental validation: For the related product influence factors,epidemic intervention factors,and other unknown random nonlinear influences that cannot be accurately measured in the error sequence,multi-feature GRU models,epidemic intervention models,and EEMD-Prophet models that can better fit nonlinear features are established to study the factors mentioned above.Meanwhile,the prediction errors of these models are used as inputs for training an RBF neural network to determine the influence weights of the three factors.Finally,a SARIMA-RBF combined model is constructed to predict future egg prices,combining epidemic intervention and multiple features.Experimental results show that the prediction accuracy of this combined model is significantly higher than that of a single prediction model.Developing an egg price fluctuation trend prediction system: The overall demand of the prediction system is analyzed,the technical framework is determined,and a visualization prediction system for egg price fluctuation trends that combines multi-feature and intervention analysis is designed and implemented.