1. College of Chemistry and Chemical Engineering, Qinghai Normal University, Xining 810016, China
2. College of Chemistry and Chemical Engineering, Qinghai Minzu University, Xining 810016, China
3. Qinghai Key Laboratory of Advanced Technology and Application of Environmentally Functional Materials, Xining 810016, China
Abstract：Chinese Huzhu Qingke Liquor is a protected geographical indication product, and it is of great significance for its accurate evaluation and classification. Due to the advantages of ultraviolet (UV) and near-infrared (NIR) spectroscopy, such as fast, accurate, non-destructive detection and no sample pretreatment, are widely used in food and other fields. In this study, a fast, nondestructive, and efficient discriminative classification model for Huzhu Qingke Liquor was established based on UV, NIR, and UV-NIR intermediate data fusion spectroscopy (UV-NIR) combined with theback-propagation neural network (BPNN) method. Since the unoptimized spectra are affected by noise and baseline drift due to the superimposed interference of spectral eigenpeaks, the spectra are denoised using four preprocessing methods, namely, standard normal variable transform (SNV), Savitzky-Golay smoothing (SG), first-order derivative (1D) and second-order derivative (2D). Further, relative to a single spectrum, the fused spectrum can complement the diversified spectroscopic information and improve the performance of the classification model, so the feature variables are selected by five variable screening methods, namely, competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA), principal component analysis (PCA), variable projection importance analysis (VIP), and variable combinatorial clustering analysis (VCPA) to achieve the optimization of model performance and the purpose of fusing the effective information of two spectra. Finally, the best method for establishing the BPNN model for single and fused spectra was selected. The results show that the classification model established by selecting 30 feature variables by SPA after SNV preprocessing for UV spectra has the best recognition effect, with a classification accuracy of 100%. The MSE value, R2P, R(Train), R(Validation), R(Test) and R(All) were 0.018 0, 1, 0.928 3, 0.958 7, 0.913 0, and 0.929 7, respectively; PCA selected the NIR and UV-NIR after SG preprocessing with 84 and 106 The classification model built by feature variables had the best recognition effect, and the NIR spectral classification accuracy was 100%, with MSE value, R2P, R(Train), R(Validation), R(Test)and R(All)of 0, 1.000, 1.000, 1.000，1.000 and 1.000. respectively, UV-NIR spectral classification accuracy was 100%, MSE, R2P, R(Train), R(Validation), R(Test), and R(All) were 0.005 7, 1.000, 1.000, 0.987 1, 0.991 3 and 0.996 4, respectively; the fusion spectra can significantly improve the predictive ability and robustness of the classification model compared with the single-spectrum modeling, thus realizing the rapid and non-destructive analysis of Huzhu Qingke Liquor.
[1] McGovern P E, Zhang J, Tang J, et al. Proceedings of the National Academy of Sciences, 2004, 101(51): 17593.
[2] ZHOU Jun, YANG Yang, YAO Yao, et al(周 军, 杨 洋, 姚 瑶, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2022, 42(3): 764.
[3] Ding X, Wu C, Huang J, et al. Journal of Food Science, 2015, 80(11): C2373.
[4] Yan S, Wang S, Wei G, et al. Journal of the Institute of Brewing, 2015, 121(1): 145.
[5] Fan Q, Wang X, Zhao Y, et al. Flavour and Fragrance Journal, 2019, 34(6): 514.
[6] Cheng P, Fan W, Xu Y. Food Research International, 2013, 54(2): 1753.
[7] Jia W, Fan Z, Du A, et al. Food Chemistry, 2020, 324: 126899.
[8] Niu Y, Zhu Q, Xiao Z. Food Research International, 2020, 131: 108986.
[9] Yan Y, Chen S, Nie Y, et al. Food Research International, 2020, 131: 109043.
[10] HE Miao(何 苗). Master Dissertation(硕士论文). Chongqing University(重庆大学), 2022.
[11] ZHOU Rui, CHEN Xiao-ming, ZHANG Li-li, et al(周 瑞, 陈晓明, 张莉丽, 等). Food and Fermentation Industry(食品与发酵工业), 2023, 49(5): 88.
[12] Linzhi D, Yongrong S, Xin W, et al. Journal of Food Composition and Analysis, 2023, 123.
[13] ZHANG Zheng-yong, SONG Chao, SHA Min, et al(张正勇, 宋 超, 沙 敏, 等). Brewing Technology(酿酒科技), 2016, (11): 20.
[14] LIU Ming-kun, HAN Qi-ying, YAN Chang-jie, et al(刘明坤, 韩奇颖, 闫长杰, 等). Brewing Technology(酿酒科技), 2022, (3): 116.
[15] CHU Xiao-li, CHEN Pu, LI Jing-yan, et al(褚小立, 陈 瀑, 李敬岩, 等). Journal of Analytical Testing(分析测试学报), 2020, 39(10): 1181.
[16] YU Si-ming, YAO Yan, LIU Ying, et al(俞思名, 姚 燕, 刘 颖, 等). Journal of China University of Weights and Measures(中国计量大学学报), 2024, 35(1): 28.
[17] DONG Xin-luo, LIU Jian-xue, HAN Si-hai, et al(董新罗, 刘建学, 韩四海, 等). Journal of Analytical Testing(分析测试学报), 2020, 39(11): 1427.
[18] HUANG Jun-shi, WANG Dong-xin, XIONG Ai-hua, et al(黄俊仕, 王冬欣, 熊爱华, 等). Journal of Jiangxi Agricultural University(江西农业大学学报), 2020, 42(6): 1270.
[19] Cai J J. Journal of Heredity, 2008, 99(4): 438.
[20] Li H, Liang Y, Xu Q, et al. Analytica Chimica Acta, 2009, 648(1): 77.
[21] Qu J, Sun D, Pu H. Water, Air, & Soil Pollution, 2017, 228(5): .
[22] Mário César Ugulino Araújo, Saldanha T C B, Galvo R K H, et al. Chemometrics & Intelligent Laboratory Systems, 2001, 57: 65.
[23] Jiao Z, Zhong Y W, Quan M Y, et al. Microchemical Journal, 2021, 160(PB): 105662.
[24] Oussama A，Elabadi F，Platikanov S，et al. Journal of the American Oil Chemists' Society, 2012, 89(10): 1807.
[25] Wold S, Johansson E. 3D QSAR in Drug Design, 1993, 1: 523.
[26] Jiang H, Xu W, Ding Y, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2020, 228(C): 117781.
[27] Yun Y, Wang W, Deng B, et al. Analytica Chimica Acta, 2015, 86214.
[28] LI Na, CHEN Wei, ZHANG Jie, et al(李 娜, 程 伟, 张 杰, 等). Brew Wine(酿酒), 2018, 45(6): 63.
