Study On The Influencing Factors Of Corn Price And Its Forecasting Method

China is a predominantly agricultural country.The healthy and sustainable development of the corn industry is an essential guarantee for the stable development of grain supply,feed consumption market,and corn processing industry,as corn production and consumption are massive and widely used.The National Bureau of Statistics data show that in 2022 China’s total grain production of 687 million tons,of which corn production amounted to 277 million tons.Studying the influencing factors and forecasting methods of corn prices is of great theoretical and practical importance.This paper takes corn as the research object,selects the national yellow corn second class market price from 2015-2021 in the WIND database,analyzes the influencing factors of corn price and establishes the forecasting model,and develops the corn price analysis and forecasting system.The details are as follows:(1)A vector autoregressive model(VAR)was developed with monthly data on corn prices and 14 influencing factors.As the confidence level increases from 1% to 10%,corn starch,pork production,corn imports,crude oil prices,pork prices,the Baidu index,and money supply are the Granger causes of corn prices.The results of the corresponding impulse analysis indicate that soybean price,corn starch price,pork price,Baidu search index,pork production,wheat production,and corn imports have positive shocks on corn price.In contrast,wheat prices and money supply negatively affect corn prices.The contribution of each factor to the change in corn price was analyzed through variance decomposition.The results indicate that the contribution of soybean price,crude oil price,pork price,pork production,and Baidu search index was more significant except for corn price.(2)A study of corn price forecasting models.Adaptive-LASSO was used for feature selection,and four feature variables were identified: soybean price,corn starch price,planted area,and money supply.The LSTM model for corn price forecasting was constructed using all the explanatory variables,the four characteristic variables selected by Adaptive-LASSO,and the explanatory variables whose parameters showed significance in the VAR model as inputs to compare the forecasting effects on corn prices.The results show that the feature variables obtained by Adaptive-LASSO reduce the redundancy compared with the original data and have better prediction results than the explanatory variables obtained by the VAR model.Finally,based on the Adaptive-LASSO-LSTM,the hyperparameters of the LSTM are optimized by the heuristic optimization algorithm SSA,and the new algorithm has a better optimization effect than the original model and further improves the prediction fit.(3)Design and implementation of corn price analysis and prediction system.Based on the analysis of corn price influencing factors and price forecasting,we further developed the system requirement analysis,and combined with the experience of big data construction,we completed the design and development of corn price forecasting service system by using data collection,data storage,data analysis,data visualization and other technologies.The corn price forecasting system can provide the query of corn price historical data and the visualization service of corn future forecast price.It can enable corn-related industry practitioners to pay attention to the trend of corn prices,adjust corn planting and consumption strategies promptly,and promote the healthy development of the corn industry.