摘要
基于生猪价格影响的供给、需求和市场外部因素,考虑其与生猪价格之间复杂的非线性关系,使用自适应LASSO回归与神经网络用于预警生猪价格波动,旨在监测生猪价格波动风险。首先利用自适应LASSO回归确定影响生猪价格的主要指标变量,将筛选出的各指标和生猪价格分别作为广义回归神经网络GRNN模型的输入和输出建立预警模型,利用该模型对安徽省2009年—2019年的生猪价格及其影响因素分析得出玉米价格、鸡蛋价格等指标可以作为生猪价格警情指标;GRNN神经网络模型输出的生猪价格在收敛速度优于BP和RBF等其它神经网络模型并且其网络输出的相对误差仅为0.056%。
Based on the supply,demand and market external factors affected by the price of hogs,considering the complex non-linear relationship between it and the price of hogs,using adaptive LASSO regression and neural network to warn hog price fluctuations to monitor the risk of hog price fluctuations.First of all,this paper uses adaptive LASSO regression to determine the main indicator variables that affect the price of hogs.The selected indicators and live hog price are used as the input and output of the GRNN neural network to establish an early warning model.The model is used to analyze the hog in Anhui Province from 2009 to 2019.The analysis of price and its factors shows that indicators such as corn price and egg price can be used as hog price warning indicators;the price of hog output by the GRNN neural network model is faster than other neural network models such as BP and RBF,and the network output’s relatively error is only 0.056%.
作者
程硕
程建华
CHENG Shuo;CHENG Jian-hua(School of Economics,Anhui University,Hefei 230601,China)
出处
《合肥学院学报（综合版）》
2020年第5期50-58,共9页
Journal of Hefei University：Comprehensive ED
基金
安徽省社会科学基金项目(AHSKF2019D019)资助
