Spectroscopy and Spectral Analysis, Volume. 42, Issue 9, 2947(2022)
A Classification Method of Coal and Gangue Based on XGBoost and Visible-Near Infrared Spectroscopy
Figures & Tables(11)
Fig. 1. Coal samples (a, c, e) and gangue samples (b, d, f) from different coal mines
(a), (b): Ximing coal mine; (c), (d): Shenmu coal mine; (e), (f): Balongtu coal mine
Fig. 3. Spectra of coal and gangue in Ⅰ: (a)(b), Ⅱ: (c)(d), Ⅲ: (e)(f) mines after pretreatment
Table 1. Samples information
View tableView in ArticleTable 1. Samples information
产地及煤矿 样品类别
及编号外观特征 样品
数量采集光
谱数量山西西铭(Ⅰ) 煤(Ⅰ.1) 黑色, 有光泽 93 169 岩(Ⅰ.2) 黑色, 无光泽 91 169 陕西神木(Ⅱ) 煤(Ⅱ.1) 黑色, 有光泽 20 39 岩(Ⅱ.2) 黑褐色, 无光泽 18 33 内蒙古巴隆图(Ⅲ) 煤(Ⅲ.1) 黑褐色, 无光泽 26 49 岩(Ⅲ.2) 灰白色, 无光泽 23 43
Table 2. Comparison of different classification models based on the full-band spectra
View tableView in ArticleTable 2. Comparison of different classification models based on the full-band spectra
Sample
originNumber of
variablesModel ACC10 ACC AUC Ⅰ 1 000 KNN 0.948 5 0.941 1 0.941 5 RF 0.953 0 0.960 7 0.961 1 SVM 0.944 3 0.960 7 0.961 1 XGBoost 0.957 2 0.970 5 0.971 6
Table 3. The prediction results of different classification models based on characteristic wavelengths
View tableView in ArticleTable 3. The prediction results of different classification models based on characteristic wavelengths
Sample
originVariable selection
methodsNumber of
variablesModel ACC10 ACC AUC Ⅰ RFE 9 KNN 0.948 3 0.950 9 0.950 9 RF 0.948 3 0.950 9 0.952 0 SVM 0.956 8 0.960 7 0.952 0 XGBoost 0.965 7 0.980 3 0.980 3 SPA 5 KNN 0.953 0 0.950 9 0.951 2 RF 0.953 0 0.960 7 0.960 7 SVM 0.965 7 0.960 7 0.961 1 XGBoost 0.957 2 0.960 7 0.960 7 CARS 61 KNN 0.957 0 0.950 9 0.951 2 RF 0.948 9 0.950 9 0.950 9 SVM 0.978 6 0.980 3 0.980 3 XGBoost 0.961 5 0.960 7 0.960 7
Table 4. The model prediction results of the test group
View tableView in ArticleTable 4. The model prediction results of the test group
Model Sample
originVariable selection
methodsNumber of
variablesACC10 ACC AUC XGBoost Ⅱ RFE 1 000 0.955 0 1.000 0 1.000 0 3 0.975 0 1.000 0 1.000 0 Ⅲ RFE 1 000 0.940 4 0.964 2 0.968 7 7 0.954 7 1.000 0 1.000 0
Tools
Get Citation
Copy Citation Text
Rui LI, Bo LI, Xue-wen WANG, Tao LIU, Lian-jie LI, Shu-xiang FAN. A Classification Method of Coal and Gangue Based on XGBoost and Visible-Near Infrared Spectroscopy[J]. Spectroscopy and Spectral Analysis, 2022, 42(9): 2947
Paper Information
Category: Research Articles
Received: Oct. 19, 2021
Accepted: Apr. 4, 2022
Published Online: Nov. 17, 2022
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20