基于EEMD-LSTM的农产品价格短期预测模型研究

本文将集成经验模态分解(EEMD)方法与长短期记忆网络(LSTM)相结合，构建了一个端到端的农产品价格短期预测模型。首先，对原始价格信号进行EEMD分解，得到若干IMF子序列和一个残差序列；然后，运用Fine-to-coarse高低频重构方法对IMF子序列进行高频—低频重构；最后将原始价格序列、高频项、低频项和残差趋势项作为特征，输入到LSTM网络进行训练得到预测模型。本文以广州市江南农副产品市场的富士苹果日价格为例进行实证分析，结果表明，本文提出的EEMD-LSTM模型在农产品价格短期预测问题上具有一定的性能优势。

Research on Short-Term Forecast Model of Agricultural Product Price Based on EEMD-LSTM

Agricultural produces price forecasting is one of the main research issues in the field of agricultural economy, and it is of great interest to many researchers and institutions.A novel method is proposed in this paper for agricultural produces price forecasting based on Ensemble Empirical Mode Decomposition (EEMD) method and Long Short-Term Memory (LSTM) network. First, the original price series is decomposed by EEMD to several IMF subsequences and one residual sequence. Then, the fine-to-coarse method is used to reconstruct the IMF subsequences into high frequency item and low frequency item. Finally, the original price series, high frequency item, low-frequency item and residual sequence are taken as input features to train the LSTM network, so as to construct an EEMD-LSTM combined forecasting model.