BP神经网络与多元线性回归的净水装置脱氮预测比较

李金泽; 叶春; 李春华; 罗艳丽; 魏伟伟

doi:10.16803/j.cnki.issn.1004-6216.2018.05.010

环境保护科学

, , , ,

李金泽, 叶春, 李春华, 罗艳丽, 魏伟伟. BP神经网络与多元线性回归的净水装置脱氮预测比较[J]. 环境保护科学, 2018, (5): 52-55. doi: 10.16803/j.cnki.issn.1004-6216.2018.05.010

引用本文:

Li Jinze, Ye Chun, Li Chunhua, Luo Yanli, Wei Weiwei. Comparison of BP Neural Network and Multiple Linear Regression in the Prediction of Nitrogen Removal Efficiency in a Surface Water Purification Device[J]. Environmental Protection Science, 2018, (5): 52-55. doi: 10.16803/j.cnki.issn.1004-6216.2018.05.010

Citation:

Li Jinze, Ye Chun, Li Chunhua, Luo Yanli, Wei Weiwei. Comparison of BP Neural Network and Multiple Linear Regression in the Prediction of Nitrogen Removal Efficiency in a Surface Water Purification Device[J]. Environmental Protection Science, 2018, (5): 52-55. doi: 10.16803/j.cnki.issn.1004-6216.2018.05.010

BP神经网络与多元线性回归的净水装置脱氮预测比较

1. 新疆农业大学草业与环境科学学院, 新疆乌鲁木齐 830052;
2. 燕京理工学院建筑学院, 河北廊坊 065201;
3. 中国环境科学研究院, 北京 100012

作者简介: 李金泽(1994-),男,硕士、助教。研究方向:市政工程和装置相关模型理论研究与应用。E-mail:LJZ_zky@163.com

基金项目:
国家水体污染控制与治理科技重大专项(2017ZX07203-005)基金资助
中图分类号: X524

Comparison of BP Neural Network and Multiple Linear Regression in the Prediction of Nitrogen Removal Efficiency in a Surface Water Purification Device

1. College of Grassland and Environmental Science, Xinjiang Agricultural University, Urumqi 830052, China;
2. School of Architecture, Yanching Institute of Technology, Langfang 065201, China;
3. Chinese Research Academy of Environmental Sciences, Beijing 100012, China
Fund Project:

摘要: 地表水净化装置在实际运行过程中,其最终出水水质会受多种因素影响制约,且该过程具有非线性、时变性与随机性,无法进一步分析该装置水质净化效果。对此,文章分别提出基于BP神经网络和多元线性回归对地表水净化装置水净化效果进行预测。结果表明,BP神经网络与多元回归预测TN准确度都较为良好,通过将拟合曲线得到的TN预测值与TN实际值进行绝对误差和相对误差分析。说明BP神经网络在TN预测上优于多元线性回归,更适用于该地表水净化装置脱氮过程预测的拟合计算。
- BP神经网络 /
- 多元线性回归 /
- 绝对误差 /
- 相对误差
Abstract: In the actual operation process, water quality of a surface water purification device will be restricted by many factors, and the process is nonlinear, time varying and stochastic, so it is impossible to analyze the effects of water purification device. Therefore, in this paper, a prediction model of water purification effects based on BP neural network and multiple linear regression is presented. The results show that the predicted results of BP neural network and multiple linear regression are good. Finally, fitted curve is used to obtain the TN predicted value and TN actual value for analysis of the absolute error and the relative error, indicating that the BP neural network is superior to the multiple linear regression in TN prediction and more suitable for fitting calculation of predicting the nitrogen removal process of the surface water purification device.
- BP Neural Network /
- Multiple Linear Regression /
- Absolute Error /
- Relative Error

点击查看大图

计量

文章访问数: 911
HTML全文浏览数: 911
PDF下载数: 3
施引文献: 0

引用本文:

Citation:

Li Jinze, Ye Chun, Li Chunhua, Luo Yanli, Wei Weiwei. Comparison of BP Neural Network and Multiple Linear Regression in the Prediction of Nitrogen Removal Efficiency in a Surface Water Purification Device[J]. Environmental Protection Science, 2018, (5): 52-55. doi: 10.16803/j.cnki.issn.1004-6216.2018.05.010

BP神经网络与多元线性回归的净水装置脱氮预测比较

作者简介: 李金泽(1994-),男,硕士、助教。研究方向:市政工程和装置相关模型理论研究与应用。E-mail:LJZ_zky@163.com
1. 新疆农业大学草业与环境科学学院, 新疆乌鲁木齐 830052;

2. 燕京理工学院建筑学院, 河北廊坊 065201;

3. 中国环境科学研究院, 北京 100012

收稿日期: 2018-04-27

基金项目:

国家水体污染控制与治理科技重大专项(2017ZX07203-005)基金资助

关键词:

摘要: 地表水净化装置在实际运行过程中,其最终出水水质会受多种因素影响制约,且该过程具有非线性、时变性与随机性,无法进一步分析该装置水质净化效果。对此,文章分别提出基于BP神经网络和多元线性回归对地表水净化装置水净化效果进行预测。结果表明,BP神经网络与多元回归预测TN准确度都较为良好,通过将拟合曲线得到的TN预测值与TN实际值进行绝对误差和相对误差分析。说明BP神经网络在TN预测上优于多元线性回归,更适用于该地表水净化装置脱氮过程预测的拟合计算。

English Abstract

Comparison of BP Neural Network and Multiple Linear Regression in the Prediction of Nitrogen Removal Efficiency in a Surface Water Purification Device

1. College of Grassland and Environmental Science, Xinjiang Agricultural University, Urumqi 830052, China;
2. School of Architecture, Yanching Institute of Technology, Langfang 065201, China;
3. Chinese Research Academy of Environmental Sciences, Beijing 100012, China

Received Date: 2018-04-27

Fund Project:

Keywords:

Abstract: In the actual operation process, water quality of a surface water purification device will be restricted by many factors, and the process is nonlinear, time varying and stochastic, so it is impossible to analyze the effects of water purification device. Therefore, in this paper, a prediction model of water purification effects based on BP neural network and multiple linear regression is presented. The results show that the predicted results of BP neural network and multiple linear regression are good. Finally, fitted curve is used to obtain the TN predicted value and TN actual value for analysis of the absolute error and the relative error, indicating that the BP neural network is superior to the multiple linear regression in TN prediction and more suitable for fitting calculation of predicting the nitrogen removal process of the surface water purification device.