• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 10, 3189 (2021)
Guo-liang WANG1、*, Ke-qiang YU3、3;, Kai CHENG2、2;, Xin LIU2、2;, Wen-jun WANG1、1;, Hong LI2、2;, Er-hu GUO2、2;, and Zhi-wei LI1、1; *;
Author Affiliations
  • 11. College of Agricultural Engineering, Shanxi Agricultural University, Taigu 030801, China
  • 22. Millet Research Institute, Shanxi Agricultural University, Changzhi 046000, China
  • 33. College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling 712100, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)10-3189-05 Cite this Article
    Guo-liang WANG, Ke-qiang YU, Kai CHENG, Xin LIU, Wen-jun WANG, Hong LI, Er-hu GUO, Zhi-wei LI. Hyperspectral Technique Coupled With Chemometrics Methods for Predicting Alkali Spreading Value of Millet Flour[J]. Spectroscopy and Spectral Analysis, 2021, 41(10): 3189 Copy Citation Text show less
    Average spectral curves of millet flour
    Fig. 1. Average spectral curves of millet flour
    Selection of key variables using CARS algorithm(a): Changing trend of the number of sampled variables; (b): Variation of root-mean-square error of cross-validation values;(c): Regression coefficients of each variable with the increasing of sampling runs
    Fig. 2. Selection of key variables using CARS algorithm
    (a): Changing trend of the number of sampled variables; (b): Variation of root-mean-square error of cross-validation values;(c): Regression coefficients of each variable with the increasing of sampling runs
    Selection probabilities of each wavelength using RF algorithm
    Fig. 3. Selection probabilities of each wavelength using RF algorithm
    The fit of training set and prediction set pretreated by MSC
    Fig. 4. The fit of training set and prediction set pretreated by MSC
    Sample
    Number
    Minimum
    /℃
    Maximum
    /℃
    Mean
    /℃
    Standard
    deviation
    35875.2584.2579.651.98
    Table 1. Statistic results of alkali spreadingvalues in millet flour
    PretreatmentNumber of
    variables
    RcRMSECRpRMSEP
    RAW1480.730.990.770.83
    CARS160.740.960.720.93
    RF100.71.030.70.90
    Table 2. PLSR modeling results of different methods based on key wavelengths extraction
    PretreatmentRcRMSECRpRMSEP
    S-G0.711.000.740.91
    MSC0.780.850.830.83
    S-G+MSC0.730.940.711.06
    Table 3. Analysis results of PLSR models by different pretreatments
    PretreatmentNumber of
    variables
    RcRMSECRpRMSEP
    CARS100.740.990.70.83
    RF100.711.030.660.87
    Table 4. PLSR predictive modeling results of different key wavelengths extractions pretreated by MSC
    Guo-liang WANG, Ke-qiang YU, Kai CHENG, Xin LIU, Wen-jun WANG, Hong LI, Er-hu GUO, Zhi-wei LI. Hyperspectral Technique Coupled With Chemometrics Methods for Predicting Alkali Spreading Value of Millet Flour[J]. Spectroscopy and Spectral Analysis, 2021, 41(10): 3189
    Download Citation