基于不同人工神经网络的水质预测方法对比  

Comparative Study of Water Quality Prediction Methods Based on Different Artificial Neural Network

作　　者：肖明君 朱逸纯 高雯媛 曾钰 李濠 陈硕夫 刘萍[1] 黄红丽 XIAO Ming-jun;ZHU Yi-chun;GAO Wen-yuan;ZENG Yu;LI HaO_();CHEN Shuo-fu;LIU Ping;HUANG Hong-li(College of Environment and Ecology,Hunan Agricultural University,Changsha 410128,China;Ecological Environment Monitoring Center of Hunan Province,Changsha 410000,China)

机构地区：[1]湖南农业大学环境与生态学院,长沙410128 [2]湖南省生态环境监测中心,长沙410000

出　　处：《环境科学》2024年第10期5761-5767,共7页Environmental Science

基　　金：国家自然科学基金青年科学基金项目(42107308)。

摘　　要：利用现有水质数据对未来水质变化进行预测是实现区域规划与流域管理的有效工具.基于2022年4月至2022年5月湘江流域衡阳段水质监测数据,构建了反向传播神经网络(BPNN)与卷积神经网络(CNN)水质指标预测模型,并对高锰酸盐指数的预测结果进行了比较分析.数据显示,BPNN模型的预测值与真实水质情况基本吻合,但在拟合过程中出现了过拟合现象,利用粒子群算法(PSO)改进BPNN模型的参数选择方式,能够避免这一现象.而CNN模型拥有更复杂的结构、更科学的拟合方法,从而避免了模型在拟合过程中陷入局部极值,同时提高了模型预测结果的准确性.以不同模型的均方根误差(RMSE)、决定系数(R2)与平均绝对误差(MAE)作为评价参数,与传统的BPNN模型相比,PSO-BPNN模型中测试集的RMSE从0.2782 mg·L^(-1)降低到0.2109 mg·L^(-1),MAE从0.2223 mg·L^(-1)降低到0.1537 mg·L^(-1),R^(2)从0.8640提高到0.9218,PSO-BPNN模型拥有更加稳定的拟合效果.CNN模型测试集的RMSE、MAE和R^(2)分别为0.1220 mg·L^(-1)、0.0927 mg·L^(-1)和0.9705,显示CNN模型预测效果更好.The prediction of future data using existing data is an effective tool for regional planning and watershed management.The back propagation neural network(BPNN)and convolutional neural network(CNN)were used to construct a prediction model based on the water quality index of Hengyang in Xiangjiang River Basin from April to May 2022 and the results of permanganate index prediction by different models were compared.The prediction results displayed by BPNN could predict the water quality;however,overfitting occurred during the prediction.BPNN modified by particle swarm optimization(PSO)could avoid overfitting,which improved the parameter selection method of the BPNN mode.The CNN model had a better prediction effect,which had a more complex structure and a more scientific fitting method to avoid the model falling into the local extreme value during the fitting process and improve the accuracy of the model prediction results.The evaluation parameters including root-mean-square error(RMSE),coefficient of determination(R2),and mean absolute error(MAE)were used to predict the accuracy of the network.Compared with that of the traditional BPNN model,PSO-BPNN reduced the RESM of the test set from 0.2782 mg·L^(-1) to 0.2109 mg·L^(-1),reduced the MAE of the test set from 0.2223 mg·L^(-1) to 0.1537 mg·L^(-1),and increased the R^(2) of the test set from 0.8640 to 0.9218,which indicated that PSO-BPNN had more stable fitting ability.RMSE,MAE,and R^(2) of the test set in the CNN model were 0.1220 mg·L^(-1),0.0927 mg·L^(-1),and 0.9705,respectively,which showed that CNN had a better fitting and prediction effect than that of BPNN.

关 键 词：湘江流域 水质预测 机器学习 人工神经网络 模型性能 

分 类 号：X52[环境科学与工程—环境工程]