摘要
为提高池塘溶氧量预测精度并改善预测结果滞后的情况,本研究提出基于经验模态分解(empirical modedecomposition,EMD)与K-means聚类的改进长短期记忆神经网络(improvedlongshort-timememory,IL⁃STM)模型。利用皮尔森相关性分析与主成分分析结合的方法对原始数据进行特征提取,对溶氧量进行EMD分解,将选出的环境参数与溶氧量各分量一起生成样本集,并对其进行K-means聚类。针对同类中不同分解分量建立相应ILSTM预测模型,并用网格搜索、五折交叉验证与早停法进行超参数选取。对未来1 h池塘溶氧量进行预测,并与LSTM、ILSTM、LSTM-SVR、EMD-LSTM、EMD-ILSTM模型进行对比试验。结果显示,IL⁃STM与LSTM模型相比,RMSE、MAE与MAPE分别下降了50.46%、63.20%与68.96%,证明ILSTM模型能缓解传统LSTM模型预测的滞后情况。EMD-ILSTM模型与ILSTM模型相比,RMSE、MAE与MAPE分别下降了53.22%、46.74%与38.19%,证明EMD算法能提高预测精度。EMD-KILSTM模型的RMSE、MAE、MAPE分别为0.1099 mg/L、0.0749 mg/L、9.3278%,与EMD-ILSTM模型相比,分别下降了4.35%、7.42%与8.09%,证明K-means聚类能提高预测精度,并且EMD-KILSTM模型是对比模型中预测效果最好的模型。以上结果表明,EMD-KILSTM模型能从时间尺度与历史环境类别两个方面深度分析溶氧量的特征,拥有更高的预测精度与更好的泛化能力。
In order to improve the prediction accuracy of dissolved oxygen in pond,and improve the lag of prediction results,this study proposed an improved long short-term memory(ILSTM)model based on empirical mode decomposition(EMD)and K-means clustering.A combination of Pearson correlation analysis and principal component analysis was used to extract features from the original data,EMD was used to decompose dissolved oxygen,and the selected environmental parameters were combined with each component of dissolved oxygen to generate a sample set to be clustered by K-means.The corresponding IL⁃STM prediction models were established for different decomposition components in the same kind,and the hyperparameters were selected by grid search,five-fold cross-validation and early stop method.The dis⁃solved oxygen in the future 1 h pond was predicted and compared with models of LSTM,ILSTM,LSTM-SVR,EMD-LSTM,and EMD-ILSTM.The results showed that the RMSE,MAE and MAPE decreased by 50.46%,63.20%and 68.96%,respectively,compared with the LSTM model,which proved that the IL⁃STM model could alleviate the prediction lag of the traditional LSTM model.Compared with ILSTM mod⁃el,RMSE,Mae and MAPE of EMD-ILSTM model,decreased by 53.22%,46.74%and 38.19%respec⁃tively,which proved that EMD Algorithm can improve the prediction accuracy.The RMSE,MAE and MAPE of the EMD-KILSTM model were 0.1099 mg/L,0.0749 mg/L and 9.3278%,respectively,and its RMSE,MAE and MAPE decreased by 4.35%,7.42%and 8.09%,respectively,compared with the EMD-ILSTM model,which proved that K-means clustering could improve the prediction accuracy and the EMD-KILSTM model was the best one among the compared models.The above results show that the EMD-KILSTM model can deeply analyze the characteristics of dissolved oxygen from both time scale and historical environmental categories,and has higher prediction accuracy and better generalization ability,which provides scientific basis for intelligent water quality control.
作者
谢雨茜
李路
朱明
谭鹤群
李家庆
宋均琦
XIE Yuxi;LI Lu;ZHU Ming;TAN Hequn;LI Jiaqing;SONG Junqi(College of Engineering,Huazhong Agricultural University,Wuhan 430070,China;Engineering Research Center of Green Development for Conventional Aquatic Biological Industry in the Yangtze River Economic Belt,Ministry of Education,Wuhan 430070,China;Key Laboratory of Aquaculture Facilities Engineering,Ministry of Agriculture and Rural Affairs,Wuhan 430070,China)
出处
《华中农业大学学报》
CAS
CSCD
北大核心
2022年第3期200-210,共11页
Journal of Huazhong Agricultural University
基金
中央高校基本科研业务费专项(2662020SCPY003,107-11041910103)
国家自然科学基金项目(31972797)。
关键词
池塘养殖
溶解氧
长短期记忆神经网络
经验模态分解
K-MEANS聚类
预测模型
pond farming
dissolved oxygen
long short-term memory neural networks
empirical mode decomposition
K-means clustering
predicting model
