【摘要】：國內豬肉市場價格具有波動大、非線性、非平穩(wěn),且樣本量少的特點,很難進行預測。為了提高預測精度,并有效解釋價格波動的內在經(jīng)濟含義,基于集成預測思想,提出EMD-SVM集成預測模型。首先用經(jīng)驗模態(tài)分解方法(EMD)把豬肉市場月度價格分解成若干個不同尺度的,相對平穩(wěn)的本征模態(tài)分量(IMF),按照頻率高低,將各IMF分量集成為高頻部分、低頻部分和殘余項三大模塊,解決波動大、非平穩(wěn)問題。在此基礎上運用支持向量機(SVM)對3個集成模塊分別進行預測,從而解決非線性問題。為了使預測模型最優(yōu),SVM的參數(shù)用遺傳算法進行尋優(yōu)。最后對3個集成模塊的預測結果再次進行集成,重構出豬肉市場價格預測值。為了驗證模型的有效性,將EMD-SVM集成預測模型與SVM、EMD-BP、BP的預測結果進行分類比較,其RMSE、MAPE和方向性都明顯提高。
[Abstract]:Domestic pork market price has the characteristics of large fluctuation, non-linear, non-stationary, and small sample size, so it is difficult to predict. In order to improve the prediction accuracy and effectively explain the intrinsic economic meaning of price fluctuation, a EMD-SVM integrated forecasting model is proposed based on the integrated forecasting idea. Firstly, the empirical mode decomposition method (EMD) is used to decompose the monthly price of pork market into several relatively stable intrinsic modal components (IMF),) with different scales. According to the frequency level, each IMF component is integrated into a high-frequency part. Low-frequency part and residual three modules to solve large fluctuations, non-stationary problems. On this basis, support vector machine (SVM) (SVM) is used to predict the three integration modules, so as to solve the nonlinear problem. In order to optimize the prediction model, the parameters of SVM are optimized by genetic algorithm. Finally, the prediction results of the three integrated modules are integrated again, and the pork market price forecast is reconstructed. In order to verify the validity of the model, the classification and comparison of the EMD-SVM integrated prediction model with the SVM,EMD-BP,BP prediction results show that both the RMSE,MAPE and the directivity of the model are significantly improved.
【作者單位】： 華南農業(yè)大學數(shù)學與信息學院;圣點世紀科技股份有限公司;
【基金】：廣東省自然科學基金資助項目(2016A030313402) 廣東省哲學社會科學規(guī)劃資助項目(GD15CGL16)
【分類號】：F323.7;TP18
，

本文編號：2469575