基于多元非线性回归和BP神经网络模型对黄河水沙监测数据特征分析的比较
Characteristic Analysis of Water and Sediment Monitoring Data of the Yellow River Based on Multiple Nonlinear Regression and BP Neural Network Model Comparison
孔豪杰1
作者信息
- 1. 浙江工商职业技术学院,浙江宁波 315012
- 折叠
摘要
利用2016-2021年黄河水位、水流量和含沙量已有的历史数据,采用三次样条插值方法,可建立多元非线性回归和BP神经网络模型.比较两种模型的误差率,进而得到BP神经网络预测精度更高(平均误差率:0.1981).这为预测含沙量提供可靠的依据,也为监管机关制定合理有效的检测方案提供了有力的支持.
Abstract
Based on the historical data of the Yellow River water level,flow rate,and sediment concentration from 2016 to 2021,a multivariate nonlinear regression and BP neural network model can be established by using the cubic spline interpolation method.Comparing the error rates of the two models,it can be concluded that the BP neural network has higher prediction accuracy(average error rate:0.1981).This provides a reliable basis for predicting sediment concentration and also provides strong support for regulatory authorities to develop reasonable and effective detection plans.
关键词
三次样条插值/多元非线性回归/BP神经网络/误差率Key words
cubic spline interpolation/multivariate nonlinear regression/BP neural network/error rate引用本文复制引用
基金项目
浙江工商职业技术学院校级教学改革项目(2022)(jg202235)
出版年
2024
NETL
NSTL
万方数据