Spectroscopy and Spectral Analysis, Volume. 41, Issue 11, 3474(2021)
Research on Oil Identification Method Based on Three-Dimensional Fluorescence Spectroscopy Combined With Sparse Principal Component Analysis and Support Vector Machine
Figures & Tables(8)
Table 1. Fluorescence characteristic parameters of four types of petroleum
View tableView in ArticleTable 1. Fluorescence characteristic parameters of four types of petroleum
Classification Em peak range/nm Ex peak range/nm Number Relative intensity of main peak/(a.u) Diesel oil 295~360
390~405280~310
330~3402 3.302~4.874
2.907~4.713Jet fuel 280~340 270~290 1 4.377~7.028 Gasoline 280~340 270~280 1 5.844~5.952 Lubricating oil 300~365 290~310 1 3.964~5.571
Table 2. The best parameters and accuracy of SVM model
View tableView in ArticleTable 2. The best parameters and accuracy of SVM model
模型参数 训练集
准确率
/%预测集
准确率
/%惩罚参
数c核参数
gPCA-SVM 1.741 1 0.006 8 100 90 SPCA-SVM 0.189 5 9.189 6 100 95
Table 3. Classification results obtained by different algorithms (accuracy rate/%)
View tableView in ArticleTable 3. Classification results obtained by different algorithms (accuracy rate/%)
Classification Feature extraction PCA SPCA KNN 85% 90% SVM 90% 95%
Tools
Get Citation
Copy Citation Text
De-ming KONG, Hong-jie CHEN, Xiao-yu CHEN, Rui DONG, Shu-tao WANG. Research on Oil Identification Method Based on Three-Dimensional Fluorescence Spectroscopy Combined With Sparse Principal Component Analysis and Support Vector Machine[J]. Spectroscopy and Spectral Analysis, 2021, 41(11): 3474
Paper Information
Category: Orginal Article
Received: Oct. 14, 2020
Accepted: --
Published Online: Dec. 17, 2021
The Author Email: De-ming KONG (chjmessage@163.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20