摘要
针对近年频现"价高伤民,价贱伤农"的"猪周期"现象,尝试使用集成经验模态分解(EEMD)方法挖掘出"猪周期"的价格波动机制,并引入遗传算法(GA)改进支持向量机。研究结果发现,通过EEMD方法能较好地展示出"猪周期"的循环轨迹;通过对比常用的预测模型,发现基于EEMD的GASVM模型预测精测更高,是一种更具有科学性的价格预测工具。
In view of the"hog cycle"phenomenon that frequently occurs in recent years,"high-price and highrisk-to-kill farmers",this paper attempts to use the integrated empirical modality method EEMD method to excavate the"hog cycle"of pork price fluctuations as a predictive length criterion,and introduces The genetic algorithm(GA)is used to optimize the performance parameters such as the penalty parameter C,kernel function g,and loss function p of the support vector machine(SVM)to further optimize the prediction performance of the SVM.The results showed that:Digging through the EEMD method can accurately dig out the"hog cycle"of pork prices;through the comparison of commonly used prediction models,the optimization performance of the support vector machine after optimization by the genetic algorithm is optimal and robust enough.Sex,and more suitable for short"porcine cycle"predictions.The GA-SVM model proposed in this paper helps to guard against the cyclical risks of pork price volatility and is a more scientific price forecasting tool.
作者
姜百臣
冯凯杰
彭思喜
JIANG Bai-chen;FENG Kai-jie;PENG Si-xi(College of Economics&Management,South China Agricultural University,Guangzhou 510642,China)
出处
《广东农业科学》
CAS
2018年第12期158-164,共7页
Guangdong Agricultural Sciences
基金
广东省自然科学基金(2017A030313425)
广州市科技计划项目(201806030008)
关键词
猪肉价格预测
支持向量机
遗传算法
集成经验模态分解
猪周期
pork price prediction
support vector machine
genetic algorithm
integrated empirical mode decomposition
hog cycle