• Spectroscopy and Spectral Analysis
  • Vol. 40, Issue 11, 3451 (2020)
Jie-hong CHENG1、1、* and Zheng-guang CHEN1、1
Author Affiliations
  • 1[in Chinese]
  • 11. College of Electrical and Information, Heilongjiang Bayi Agricultural University, Daqing 163319, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2020)11-3451-06 Cite this Article
    Jie-hong CHENG, Zheng-guang CHEN. Wavelength Selection of Near-Infrared Spectra Based on Improved SiPLS-Random Frog Algorithm[J]. Spectroscopy and Spectral Analysis, 2020, 40(11): 3451 Copy Citation Text show less
    Original and pre-processed spectra(a): Original; (b): S-G first derivative
    Fig. 1. Original and pre-processed spectra
    (a): Original; (b): S-G first derivative
    The result of random frog
    Fig. 2. The result of random frog
    The result of Si-RF(a): Selection probability of each variable;(b): RMSEV values of each model
    Fig. 3. The result of Si-RF
    (a): Selection probability of each variable;(b): RMSEV values of each model
    Correlation between measured and predicted values of SOM obtained from different models(a): Full spectrum PLS; (b): RF-MLR; (c): Si-RF-MLR
    Fig. 4. Correlation between measured and predicted values of SOM obtained from different models
    (a): Full spectrum PLS; (b): RF-MLR; (c): Si-RF-MLR
    样本类型样本数最小值最大值均值标准差变异
    系数/%
    总体样本10842.9195.8585.4310.8212.7
    建模集8142.9195.8583.7311.7914.0
    预测集2774.9993.4690.524.304.8
    Table 1. Statistical data of soil organic matter content
    区间总数组合数所选区间RMSECV
    202[10 18]1.678
    3[10 15 19]1.595
    252[12 23]1.787
    3[12 19 24]1.575
    302[15 28]1.556
    3[12 15 28]1.466
    352[15 17]1.703
    3[17 18 32]1.583
    Table 2. Sub-interval optimization results of SiPLS
    迭代次数变量数RMSEVmin
    500130.988 4
    1 000170.818 4
    1 500240.876 5
    2 000170.981 9
    Table 3. Optimal results of different iteration times
    模型变量数RcRMSECRpRMSEP
    FULL-PLS1 0500.963 33.143 10.909 31.954 3
    RF-MLR100.982 72.172 00.935 41.627 6
    Si-RF-MLR170.993 61.327 00.984 80.818 4
    Table 4. Results of model with different wavelength selection methods
    Jie-hong CHENG, Zheng-guang CHEN. Wavelength Selection of Near-Infrared Spectra Based on Improved SiPLS-Random Frog Algorithm[J]. Spectroscopy and Spectral Analysis, 2020, 40(11): 3451
    Download Citation