Spectroscopy and Spectral Analysis, Volume. 42, Issue 8, 2353(2022)
A Comparative Study of the COD Hyperspectral Inversion Models in Water Based on the Maching Learning
Figures & Tables(11)
Fig. 2. Spectral profiles of water samples after preprocess
(a): SG smoothing; (b): MSC; (c): SG smoothing and MSC
Fig. 3. Relationship between the number of decision trees and model MSE on training sample
(a): Random forest; (b): Adaboost; (c): XGBoost
Fig. 4. Sccetterplots of XGBoost inversion model based on different preprocessing methods
(a): Original data; (b): MSC; (c): SG smoothing; (d): SG smoothing and MSC
Fig. 5. The variancecontribution rate of the first ten principal components about PCA
Table 1. Results of chemical oxgen demand (COD) statictical value of samples
View tableView in ArticleTable 1. Results of chemical oxgen demand (COD) statictical value of samples
数据集 数量 均值/
(mg·L-1)最小值/
(mg·L-1)最大值/
(mg·L-1)SD/
(mg·L-1)训练集 1 238 134.8 54.2 216.7 25.0 预测集 310 133.8 74.8 258.6 24.3 总样本集 1 548 134.6 54.2 258.6 24.8
Table 2. The results of machine learning model based on orginal data
View tableView in ArticleTable 2. The results of machine learning model based on orginal data
机器学习
模型测试集指标 训练时间
/sR2 RMSE/(mg·L-1) RPD 多元线性 0.48 18.5 1.3 1.7 随机森林 0.85 10.0 2.4 146 AdaBoost 0.87 7.6 3.2 194 XGBoost 0.91 7.1 3.4 98.9
Table 3. The results of machine learning model based on data processed by SG smoothing
View tableView in ArticleTable 3. The results of machine learning model based on data processed by SG smoothing
机器学习
模型测试集指标 训练时间
/sR2 RMSE/(mg·L-1) RPD 多元线性 0.3 28.3 0.86 2 随机森林 0.84 10.0 2.4 188 AdaBoost 0.88 8.3 2.9 165 XGBoost 0.91 7.0 3.5 70
Table 4. The results of machine learning model based on data processed by MSC
View tableView in ArticleTable 4. The results of machine learning model based on data processed by MSC
机器学习
模型测试集指标 训练时间
/sR2 RMSE/(mg·L-1) RPD 多元线性 0.21 29.1 0.84 2.2 随机森林 0.82 10.3 2.4 124 AdaBoost 0.83 10.3 2.4 178 XGBoost 0.90 7.3 3.3 72
Table 5. The results of machine learning model based on data processed by SG smoothing and MSC
View tableView in ArticleTable 5. The results of machine learning model based on data processed by SG smoothing and MSC
机器学习
模型测试集指标 训练时间
/sR2 RMSE/(mg·L-1) RPD 多元线性 0.21 29.1 0.84 2.2 随机森林 0.86 9.1 2.7 145 AdaBoost 0.87 8.9 2.7 163 XGBoost 0.92 7.1 3.4 72
Table 6. The result of the XGBoost model built based on the PCA method
View tableView in ArticleTable 6. The result of the XGBoost model built based on the PCA method
特征提取方法 测试集指标 训练时间
/sR2 RMSE/(mg·L-1) RPD 未经过PCA处理 0.92 7.1 3.4 72 PCA处理 0.93 6.4 3.8 2.9
Tools
Get Citation
Copy Citation Text
Chun-ling WANG, Kai-yuan SHI, Xing MING, Mao-qin CONG, Xin-yue LIU, Wen-ji GUO. A Comparative Study of the COD Hyperspectral Inversion Models in Water Based on the Maching Learning[J]. Spectroscopy and Spectral Analysis, 2022, 42(8): 2353
Paper Information
Category: Orginal Article
Received: Jun. 15, 2021
Accepted: --
Published Online: Mar. 17, 2025
The Author Email: WANG Chun-ling (wangchl@bjfu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20