改进的BP神经网络模型在辽宁中部河流水质预测中的应用研究  

Application of Improved BP Neural Network Model to Prediction of River Water Quality in Middle Part of Liaoning Province Author

作　　者：郑鹏 Zheng Peng

机构地区：[1]辽宁省葫芦岛水文局,辽宁葫芦岛122500

出　　处：《吉林水利》2017年第2期18-21,共4页Jilin Water Resources

摘　　要：本文采用改进的BP神经网络模型对辽宁中部某河流水质进行预测。结果表明:改进的BP神经网络模型引入横向和纵向伸缩修正系数对模型梯度函数进行改进,提高传统BP模型收敛和计算精度。在区域河流水质预测精度明显好于传统模型,预测的河流水质总氮指标值相对误差均值明显减少,月尺度过程相关系数有较大提高。In this paper,an improved BP neural network model is used to forecast the water quality of a river in central Liaoning.The results show that the modified BP neural network model introduces the transverse and longitudinal dilation correction coefficients to improve the model gradient function and improve the convergence and precision of the traditional BP model.The prediction accuracy of regional river water quality is obviously better than that of traditional model.The relative error mean value of total nitrogen index of river water quality is decreased by 12.7%,and the correlation coefficient of monthly scale process is increased by 0.25.The research results can provide reference value for regional river water quality prediction methods.

关 键 词：改进的BP神经网络模型 水平和横向伸缩修正系数 水质预测 辽宁中部河流 

分 类 号：P342[天文地球—水文科学]