Food and Fermentation Industries >

2025 , Vol. 51 >Issue 24: 359 - 367

DOI: https://doi.org/10.13995/j.cnki.11-1802/ts.042213

Detection of base liquor grade of Nongxiangxing Baijiu by near-infrared spectroscopy based on machine learning hybrid feature selection method

  • CAO Mengde ,
  • WANG Yibo ,
  • ZHANG Guiyu ,
  • HUANG Zhichao ,
  • ZENG Xianglin ,
  • PENG Yingjie
Expand
  • 1(School of Automation & Information Engineering, Sichuan University of Science & Engineering, Yibin 644000, China)
    2(Artificial Intelligence Key Laboratory of Sichuan Province, Sichuan University of Science & Engineering, Yibin 644000, China)
    3(Liquor Making Biological Technology and Application of Key Laboratory of Sichuan Province, Yibin 644000, China)
    4(Chengdu Mingkuo Digital Intelligence Company Limited, Chengdu 610000, China)
    5(Key Laboratory of Enterprise Informatization and Internet of Things Measurement and Control Technology, Colleges and Universities of Sichuan Province, Yibin 644000, China)
    6(School of Computer Science and Engineering, Sichuan University of Science & Engineering, Yibin 644000, China)

Received date: 2025-01-21

  Revised date: 2025-06-16

  Online published: 2026-01-12

Fold

Abstract

During the grading process of Chinese liquor base liquor, sensory evaluation methods are highly subjective, while traditional methods are inefficient and prone to contaminating the samples.Therefore, it is of great significance to predict its grade quickly and accurately.This study focuses on the base liquor of strong-flavor Chinese liquor and uses near-infrared spectroscopy to establish a grade discrimination model, which is then tested for effectiveness.The method first combines cross-validation recursive feature elimination with machine learning techniques.This approach is used to measure the importance of near-infrared spectral features and perform initial extraction.Next, a forward stepwise greedy algorithm is used to determine the optimal feature subset for base liquor grade prediction.Finally, a classification boosting model is built for prediction.The performance of five machine learning feature extraction methods (multinomial logistic regression, decision tree, random forest, gradient boosting machine, and extreme gradient boosting) is compared using accuracy, recall, F1 score, confusion matrix, and receiver operating characteristic curve.The results show that the combination of cross-validation recursive feature elimination and extreme gradient boosting yields the best feature extraction performance.The final classification accuracy reaches 94.24%, an improvement of 14.14% compared to the method without feature extraction.This method demonstrates good reproducibility and strong interpretability, providing a new approach for the rapid detection of Chinese liquor base liquor grades.

Key words： near-infrared spectroscopy; base liquor grade; feature selection; machine learning; recursive feature elimination

Cite this article

CAO Mengde , WANG Yibo , ZHANG Guiyu , HUANG Zhichao , ZENG Xianglin , PENG Yingjie . Detection of base liquor grade of Nongxiangxing Baijiu by near-infrared spectroscopy based on machine learning hybrid feature selection method[J]. Food and Fermentation Industries, 2025 , 51(24) : 359 -367 . DOI: 10.13995/j.cnki.11-1802/ts.042213

References

[1] QIAN W, LU Z M, CHAI L J, et al.Cooperation within the microbial consortia of fermented grains and pit mud drives organic acid synthesis in strong-flavor Baijiu production[J].Food Research International, 2021, 147:110449.
[2] LI H H, ZHANG X, GAO X J, et al.Comparison of the aroma-active compounds and sensory characteristics of different grades of light-flavor Baijiu[J].Foods, 2023, 12(6):1238.
[3] WANG Z, WANG Y, ZHU T T, et al.Characterization of the key odorants and their content variation in Niulanshan Baijiu with different storage years using flavor sensory omics analysis[J].Food Chemistry, 2022, 376:131851.
[4] WU Z Y, QIN D, DUAN J W, et al.Characterization of benzenemethanethiol in sesame-flavour baijiu by high-performance liquid chromatography-mass spectrometry and sensory science[J].Food Chemistry, 2021, 364:130345.
[5] HE F, DUAN J W, ZHAO J W, et al.Different distillation stages Baijiu classification by temperature-programmed headspace-gas chromatography-ion mobility spectrometry and gas chromatography-olfactometry-mass spectrometry combined with chemometric strategies[J].Food Chemistry, 2021, 365:130430.
[6] LUO W, TIAN P, FAN G Z, et al.Non-destructive determination of four tea polyphenols in fresh tea using visible and near-infrared spectroscopy[J].Infrared Physics & Technology, 2022, 123:104037.
[7] ZHONG J, CHEN J M, YAO L J, et al.Discriminant analysis of liquor brands based on moving-window waveband screening using near-infrared spectroscopy[J].American Journal of Analytical Chemistry, 2018, 9(3):124-133.
[8] YANG B, YAO L J, PAN T.Near-infrared spectroscopy combined with partial least squares discriminant analysis applied to identification of liquor brands[J].Engineering, 2017, 9(2):181-189.
[9] 宗绪岩, 彭厚博, 吴键航, 等.化学计量学结合NIR对浓香型白酒年份、等级的研究[J].包装与食品机械, 2022, 40(2):87-94.
ZONG X Y, PENG H B, WU J H, et al.Study on the year and grade of Luzhou-flavor liquor by chemometrics and NIR[J].Packaging and Food Machinery, 2022, 40(2):87-94.
[10] ZHANG G Y, TUO X G, PENG Y J, et al.A rapid nondestructive detection method for liquor quality analysis using NIR spectroscopy and pattern recognition[J].Applied Sciences, 2024, 14(11):4392.
[11] FU J S, YU H D, CHEN Z, et al.A review on hybrid strategy-based wavelength selection methods in analysis of near-infrared spectral data[J].Infrared Physics & Technology, 2022, 125:104231.
[12] CHEN Y, MA H, ZHANG Q, et al.Comparison of several variable selection methods for quantitative analysis and monitoring of the Yangxinshi tablet process using near-infrared spectroscopy[J].Infrared Physics & Technology, 2020, 105:103188.
[13] REN G X, NING J M, ZHANG Z Z.Multi-variable selection strategy based on near-infrared spectra for the rapid description of dianhong black tea quality[J].Spectrochimica Acta Part A:molecular and Biomolecular Spectroscopy, 2021, 245:118918.
[14] 谢玮萍, 徐将, 魏海波, 等.碾磨程度对LIBS鉴别稻米产地准确度的影响[J].江西农业大学学报, 2025, 47(2):478-485.
XIE W P, XU J, WEI H B, et al.Evaluating impact of milling degree on rice origin identification by laser-induced breakdown spectroscopy[J].Acta Agriculturae Universitatis Jiangxiensis (Natural Sciences Edition), 2025, 47(2):478-485.
[15] 梅轩铭, 胡耀华, 张浩天, 等.基于高光谱成像技术判别马铃薯叶片干旱状态[J].干旱地区农业研究, 2024, 42(5):246-254.
MEI X M, HU Y H, ZHANG H T, et al.Evaluation of drought status of potato leaves based on hyperspectral imaging[J].Agricultural Research in the Arid Areas, 2024, 42(5):246-254.
[16] ZHENG R Y, JIA Y Y, ULLAGADDI C, et al.Optimizing feature selection with gradient boosting machines in PLS regression for predicting moisture and protein in multi-country corn kernels via NIR spectroscopy[J].Food Chemistry, 2024, 456:140062.
[17] ZHANG X, GAO Z M, YANG Y L, et al.Rapid identification of the storage age of dried tangerine peel using a hand-held near infrared spectrometer and machine learning[J].Journal of Near Infrared Spectroscopy, 2022, 30(1):31-39.
[18] TEKLER Z D, LEI Y, PENG Y Z, et al.A hybrid active learning framework for personal thermal comfort models[J].Building and Environment, 2023, 234:110148.
[19] WANG Z H, CAI Y D, LIU D M, et al.Intelligent classification of coal structure using multinomial logistic regression, random forest and fully connected neural network with multisource geophysical logging data[J].International Journal of Coal Geology, 2023, 268:104208.
[20] XU Q, LI X Y, JIANG S Z, et al.Features selection for recognition of severe slugging in a long pipeline with an S-shaped riser by decision tree[J].Flow Measurement and Instrumentation, 2024, 96:102537.
[21] TAN X Z, LI Y, WU X X, et al.Identification of the key driving factors of flash flood based on different feature selection techniques coupled with random forest method[J].Journal of Hydrology:Regional Studies, 2024, 51:101624.
[22] RUFINO J, RAMÍREZ J M, AGUILAR J, et al.Performance and explainability of feature selection-boosted tree-based classifiers for COVID-19 detection[J].Heliyon, 2023, 10(1):e23219.
[23] PROKHORENKOVA L, GUSEV G, VOROBEV A, et al. CatBoost: Unbiased Boosting With Categorical Features. Proceedings of the 32nd International Conference on Neural Information Processing Systems. ACM, 2018:6639-6649.
[24] 张劲松, 周迪, 陈明举, 等.基于稀疏主成分分析白酒GC-MS图谱分类识别技术[J].中国酿造, 2022, 41(4):216-221.
ZHANG J S, ZHOU D, CHEN M J, et al.Classification identification technology of GC-MS chromatographic data of Baijiu based on sparse principal component analysis[J].China Brewing, 2022, 41(4):216-221.
[25] YE M Q, GAO Z P, LI Z, et al.Rapid detection of volatile compounds in apple wines using FT-NIR spectroscopy[J].Food Chemistry, 2016, 190:701-708.
[26] MARTELO-VIDAL M J, VÁZQUEZ M.Evaluation of ultraviolet, visible, and near infrared spectroscopy for the analysis of wine compounds[J].Czech Journal of Food Sciences, 2014, 32(1):37-47.
[27] WORKMAN J JR.Practical Guide to Interpretive Near-Infrared Spectroscopy[M].Beijing:Chemical Industry Press, 2009:64-65.
[28] 赵玉霞, 王茹, 张世芝, 等.基于紫外-近红外融合光谱对“互助”青稞酒判别分析研究[J].食品与发酵工业, 2025, 51(4):301-308.
ZHAO Y X, WANG R, ZHANG S Z, et al.Qualitative identification of Chinese Huzhu Qingke Liquor based on ultraviolet spectroscopy combined with near infrared spectroscopy[J].Food and Fermentation Industries, 2025, 51(4):301-308.
[29] 高畅, 张宇飞, 辛颖, 等.近红外光谱技术结合波段筛选用于白酒基酒总酯定量分析[J].中国酿造, 2021, 40(4):155-158.
GAO C, ZHANG Y F, XIN Y, et al.Quantitative analysis of total esters in Baijiu base liquor by near-infrared spectroscopy combined with band selection[J].China Brewing, 2021, 40(4):155-158.
Options
Outlines

/

〈 〉

Powered by Beijing Magtech Co. Ltd,.