Excessive nitrogen, phosphorus, and permanganate in aquatic environments can lead to significant watershed pollution. Accurately predicting the levels of these indicators is crucial for effective pollution control. However, existing models often lack precision, and the selection of input factors lacks a mathematical basis. In this study, we propose a RF-BiLSTM hybrid network model focusing on the Yongjiang watershed as a case study. Leveraging the ability of RF (random forest) to extract optimal water quality index characteristics and the capacity of BiLSTM (bidirectional long-short-term memory) to capture temporal data patterns, our model employs dimensionality reduction followed by prediction to forecast TN, TP, and CODMn concentrations. Additionally, we conduct comparative analyses with benchmark models such as CNN, LSTM, BiLSTM, and RF-LSTM within the deep learning framework. Results demonstrate that our proposed model achieves lower mean absolute percentage errors (MAPE) for TN, TP, and CODMn at 4.33%, 6.781%, and 7.384%, respectively, outperforming other benchmark models. These findings indicate the high accuracy and practical utility of our predictions, offering valuable technical support for water pollution management.
Key words
water quality prediction /
feature selection /
random forest /
bidirectional long-short-term memory network /
deep learning
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
References
[1] AHMED U, MUMTAZ R, ANWAR H, et al. Efficient Water Quality Prediction Using Supervised Machine Learning[J]. Water, 2019, 11(11): 2210.
[2] TA X, WEI Y. Research on a Dissolved Oxygen Prediction Method for Recirculating Aquaculture Systems Based on a Convolution Neural Network[J]. Computers and Electronics in Agriculture, 2018, 145: 302-310.
[3] 涂吉昌, 陈超波, 王景成, 等. 基于深度学习的水质预测模型研究[J]. 自动化与仪表, 2019, 34(6): 96-100. (TU Ji-chang, CHEN Chao-bo, WANG Jing-cheng, et al. Research on Water Quality Prediction Model Based on Deep Learning[J]. Automation & Instrumentation, 2019, 34(6): 96-100.(in Chinese))
[4] WU S, LI H. Prediction of PM2.5 Concentration in Urban Agglomeration of China by Hybrid Network Model[J]. Journal of Cleaner Production, 2022, 374: 133968.
[5] 杜 洋. 基于特征选择分类和双向LSTM神经网络的钓鱼网站检测[D]. 成都: 西南交通大学, 2018. (DU Yang. Phishing Websites Detection Using Selected Features Classification and Bidirectional Long Short-term Memory Neural Networks[D].Chengdu: Southwest Jiaotong University, 2018. (in Chinese))
[6] 邹青宏. 基于多时间尺度双向LSTM网络的水质预测方法研究[D]. 重庆: 重庆大学, 2021. (ZOU Qing-hong. Study of Water Quality Prediction Based on Multi-time Scale Bidirectional LSTM[D].Chongqing: Chongqing University, 2021. (in Chinese))
[7] SIAMI-NAMINI S, TAVAKOLI N, NAMIN A S. The Performance of LSTM and BiLSTM in Forecasting Time Series[C]//2019 IEEE International Conference on Big Data (Big Data). December 9-12, 2019. Los Angeles, CA, USA. New York: IEEE Press, 2019.
[8] 吴慧英, 杨日剑, 张 颖, 等.基于PCA-SVR的池塘DO预测模型[J]. 安徽大学学报(自然科学版), 2016, 40(6): 103-108. (WU Hui-ying, YANG Ri-jian, ZHANG Ying, et al.Pond DO Prediction Model Based on PCA-SVR[J]. Journal of Anhui University (Natural Science Edition), 2016, 40(6): 103-108.(in Chinese))
[9] BREIMAN L. Random Forests[J]. Machine Learning, 2001, 45(1): 5-32.
[10]白海强, 吕保玉. 穿城河流邕江水质主成分特征分析研究[J]. 环境科学与管理, 2015, 40(1): 139-142. (BAI Hai-qiang, L Bao-yu. Application of Principal Component Analysis in Analyzing Water Quality of Urban River Yongjiang[J]. Environmental Science and Management, 2015, 40(1): 139-142.(in Chinese))
[11]ZOU Q,XIONG Q,LI Q,et al.A Water Quality Prediction Method Based on the Multi-time Scale Bidirectional Long Short-term Memory Network[J]. Environmental Science and Pollution Research,2020,27(14):16853-16864.
[12]HU Z, ZHANG Y, ZHAO Y, et al. A Water Quality Prediction Method Based on the Deep LSTM Network Considering Correlation in Smart Mariculture[J]. Sensors (Basel), 2019, 19(6): E1420.
[13]张贻婷,李天宏.基于长短时记忆神经网络的河流水质预测研究[J].环境科学与技术,2021,44(8):163-169.(ZHANG Yi-ting,LI Tian-hong.River Water Quality Prediction Based on Long Short-term Memory Neural Network[J]. Environmental Science & Technology, 2021, 44(8): 163-169.(in Chinese))
[14]陈元鹏,郧文聚,周 旭,等.基于MESMA和RF的山丘区土地利用信息分类提取[J].农业机械学报,2017,48(7):136-144.(CHEN Yuan-peng, YUN Wen-ju, ZHOU Xu, et al.Classification and Extraction of Land Use Information in Hilly Area Based on MESMA and RF Classifier[J]. Transactions of the Chinese Society for Agricultural Machinery,2017,48(7):136-144.(in Chinese))
[15]张文涛, 龚振宇, 令凡琳, 等. 基于随机森林算法的盾构改良渣土渗透系数预测及工程应用[J]. 隧道建设(中英文), 2022, 42(11): 1863-1870. (ZHANG Wen-tao, GONG Zhen-yu, LING Fan-lin, et al. Prediction and Engineering Application of Permeability Coefficients in Improved Muck of Shield Based on Random Forest Algorithm[J]. Tunnel Construction, 2022, 42(11): 1863-1870.(in Chinese))
[16]BENGIO Y, SIMARD P, FRASCONI P. Learning Long-term Dependencies with Gradient Descent is Difficult[J]. IEEE Transactions on Neural Networks, 1994, 5(2): 157-166.
[17]HOCHREITER S. The Vanishing Gradient Problem during Learning Recurrent Neural Nets and Problem Solutions[J]. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 1998, 6(2): 107-116.
[18]HOCHREITER S, SCHMIDHUBER J. Long Short-term Memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[19]PASCANU R, MIKOLOV T, BENGIO Y. On the Difficulty of Training Recurrent Neural Networks[C]//Proceedings of the 30th International Conference on International Conference on Machine Learning-Volume 28. June 16-21, 2013, Atlanta, GA, USA. 2013: III-1310–III-1318.
[20]周朝勉, 刘明萍, 王京威. 基于CNN-LSTM的水质预测模型研究[J]. 水电能源科学, 2021, 39(3): 20-23. (ZHOU Chao-mian, LIU Ming-ping, WANG Jing-wei. Research on Water Quality Prediction Model Based on CNN-LSTM[J]. Water Resources and Power, 2021, 39(3): 20-23.(in Chinese))
PDF(6941 KB)
