• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 10, 3117 (2021)
Yu-hua QIN1、*, Meng ZHANG1、1; *;, Ning YANG2、2;, and Qiu-fu SHAN3、3;
Author Affiliations
  • 11. College of Information Science and Technology, Qingdao University of Science and Technology, Qingdao 266061, China
  • 22. Qingdao Lanzhi Modern Service Industry Digital Engineering Research Center, Qingdao 266071, China
  • 33. China Tobacco Yunnan Industrial Co., Ltd., Technical Research Center, Kunming 650024, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)10-3117-06 Cite this Article
    Yu-hua QIN, Meng ZHANG, Ning YANG, Qiu-fu SHAN. Local Preserving Projection Similarity Measure Method Based on Kernel Mapping and Rank-Order Distance[J]. Spectroscopy and Spectral Analysis, 2021, 41(10): 3117 Copy Citation Text show less
    Shared neighbor information of sample points xi and xj
    Fig. 1. Shared neighbor information of sample points xi and xj
    Residual diagram with different parameter values
    Fig. 2. Residual diagram with different parameter values
    Dimensionality reduction projection of PCA, LPP, INLPP, KRLPP algorithms
    Fig. 3. Dimensionality reduction projection of PCA, LPP, INLPP, KRLPP algorithms
    质量偏差档次质量偏差描述
    0
    1轻微
    2略有
    3
    4稍大
    5较大
    Table 1. Evaluation criteria of overall quality
    算法正确识别率/%
    PCA66.2
    LPP77.6
    INLPP87.8
    KRLPP91.2
    Table 2. Comparison of the classification results of tobacco stalk position based on different dimensional reduction algorithms
    方法单料烟烟叶相
    似性
    化学成分/%感官特征质量差异
    总烟碱总糖还原糖总氮香气特征烟气特征口感特征总分
    目标烟叶-2.2425.5121.831.9735.134.415.084.5-
    替换烟叶110.82.4225.6221.941.9234.834.014.783.51.5
    PCA替换烟叶211.22.4425.6521.921.9134.734.114.683.42.0
    替换烟叶311.52.4425.6421.892.0334.833.914.783.92.5
    替换烟叶19.272.3225.5421.861.9434.934.214.883.91.0
    LPP替换烟叶29.562.3425.5621.861.9334.834.114.883.71.5
    替换烟叶39.732.3625.5721.871.9334.934.114.783.71.5
    替换烟叶14.542.2225.5221.831.9635.134.415.084.50
    KRLPP替换烟叶27.902.2025.5321.841.9535.034.415.084.40
    替换烟叶37.782.1825.5221.821.9535.034.414.084.40.5
    Table 3. Evaluation comparison of replacement tobacco and target tobacco
    方法叶组配方烟叶相
    似性
    化学成分/%感官特征质量差异
    总烟碱总糖还原糖总氮香气特征烟气特征口感特征总分
    原配方-2.2325.5321.872.0036.434.716.087.1-
    替换配方110.82.4225.6121.981.9836.134.515.786.41.5
    PCA替换配方211.22.4325.6621.951.9636.234.415.886.41.5
    替换配方311.52.4525.6521.932.0636.134.415.786.22.0
    替换配方19.272.3125.5521.891.9836.334.715.986.91.0
    LPP替换配方29.562.3425.5621.911.9736.234.715.786.61.5
    替换配方39.732.3725.5821.901.9736.334.615.886.71.0
    替换配方14.542.2225.5321.872.0036.434.716.087.10
    KRLPP替换配方27.902.2025.5421.862.0036.434.616.087.00.5
    替换配方37.782.2025.5221.871.9936.334.716.087.00.5
    Table 4. Evaluation comparison of replacement formula and original formula
    Yu-hua QIN, Meng ZHANG, Ning YANG, Qiu-fu SHAN. Local Preserving Projection Similarity Measure Method Based on Kernel Mapping and Rank-Order Distance[J]. Spectroscopy and Spectral Analysis, 2021, 41(10): 3117
    Download Citation