Research On The Price Forecast Of Pork In China Based On The Combination Model

According to the analysis of my country’s statistics department,the consumption of live pigs accounts for more than 70% of the total consumption of residents in my country’s meat consumption.The production and sales process of pork is easily affected by the market,capital and related events,resulting in frequent market price fluctuations,which in turn directly affects the daily consumption of residents and the economic development of various industries.Studying the changing laws of pork prices,predicting changes in pork prices in advance,and proposing more reasonable control measures are of the most important significance for ensuring the national economy and people’s livelihood and realizing the healthy and stable development of the pig industry chain.This paper studies the pork price prediction method of the multi-factor combination model.Different models have different emphases on data analysis and have their characteristics.The model combination can give full play to the advantages of a single model and effectively improve the prediction accuracy.At the same time,the price of pork is affected by a variety of factors,which increases the difficulty of prediction.Combining multiple factors for price prediction can more fully analyze the law of pork prices.This paper selects the relevant data on pork prices in my country by consulting relevant materials and using web crawlers to obtain data.By processing the data,it analyzes the influencing factors of pork prices and uses a multi-model combination to establish a pork price model forecast.The results show that the constructed multi-factor combination model has high accuracy,which provides theoretical and technical support for the realization of pork price prediction.The main research contents of this thesis are as follows:(1)Obtain data on pork prices,pork-related product prices,pork market supply,unexpected factors and other data on different time scales by querying relevant databases and documents,using web crawler technology,and preprocessing the data to construct pork prices.database.(2)Use the HP filtering method,Census X-13 seasonal adjustment method,Pearson correlation coefficient and other methods to analyze the characteristics of pork price changes from the aspects of the trend,periodicity,seasonality,price of related products,supply and demand relationship and the impact of epidemic diseases.After the analysis,it is finally determined to predict the pork price from the fluctuation of the pork price itself,related price factors and other factors.(3)Three methods,BP neural network model,LSTM neural network model,and SVR support vector machine regression model,were introduced to predict and analyze pork prices in different aspects.SVR(w)and stacking method BP-LSTM-SVR(s)are used for model combination,using mean absolute error(MAE),mean square error(MSE),root mean square error(RMSE)as the evaluation criteria for model performance and carrying out In comparison,the results show that although a single model can predict pork prices,a single model can only express one mapping relationship,and the prediction effect still needs to be improved.The multi-factor combination model can reduce the prediction error very well.The errors of the BP-LSTM-SVR(w)combination model are 1.26,1.69,2.85,and the errors of the BP-LSTM-SVR(s)combination model are 0.30 and 0.53,respectively.,0.28.This shows that the BP-LSTM-SVR(s)multifactor model combined with the stacking method has a better effect,and can effectively reduce the prediction error by combining the advantages of a single model,which can be used to predict pork prices.