• Acta Optica Sinica
  • Vol. 40, Issue 7, 0730002 (2020)
Mengqi Tao1、2, Jiaxiang Liu1, Yue Wu1、2, Zhiqiang Ning1、2, and Yonghua Fang1、2、*
Author Affiliations
  • 1Key Laboratory of Environmental Optics and Technology, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei, Anhui 230031, China
  • 2University of Science and Technology of China, Hefei, Anhui 230026, China
  • show less
    DOI: 10.3788/AOS202040.0730002 Cite this Article Set citation alerts
    Mengqi Tao, Jiaxiang Liu, Yue Wu, Zhiqiang Ning, Yonghua Fang. Application of XGBoost in Gas Infrared Spectral Recognition[J]. Acta Optica Sinica, 2020, 40(7): 0730002 Copy Citation Text show less
    Flow chart of XGBoost algorithm
    Fig. 1. Flow chart of XGBoost algorithm
    Schematic of XGBoost algorithm[13]
    Fig. 2. Schematic of XGBoost algorithm[13]
    Comparison before and after spectral pretreatment. (a)(b) Trichloromethane; (c)(d) paraxylene; (e)(f) tetrachloroethylene
    Fig. 3. Comparison before and after spectral pretreatment. (a)(b) Trichloromethane; (c)(d) paraxylene; (e)(f) tetrachloroethylene
    Flow chart of XGBoost model training
    Fig. 4. Flow chart of XGBoost model training
    FeatureMeaning
    WidthFull width at half maximum of characteristic peak
    KurtosisSharpness of characteristic peak
    SkewnessSymmetry of characteristic peak
    CorrelationCorrelation coefficient with standard spectrum on NIST
    SNRSignal to noise ratio of characteristic peak
    Table 1. Features for gas spectral data classification
    Gas nameTrichloromethaneParaxyleneTetrachloroethylene
    Trichloromethane887155
    Paraxylene3181411
    Tetrachloroethylene913799
    Table 2. Classification error matrix for three kinds of gases
    ModelAccuracy /%
    RF96.35
    SVM79.48
    CNN80.37
    FNN95.61
    XGBoost96.75
    Table 3. Classification accuracy for five models
    Mengqi Tao, Jiaxiang Liu, Yue Wu, Zhiqiang Ning, Yonghua Fang. Application of XGBoost in Gas Infrared Spectral Recognition[J]. Acta Optica Sinica, 2020, 40(7): 0730002
    Download Citation