食品与发酵工业 >

2026 , Vol. 52 >Issue 6: 136 - 145

DOI: https://doi.org/10.13995/j.cnki.11-1802/ts.042930

研究报告

基于机器学习的房县黄酒新鲜度预测模型研究

  • 何文杰 ,
  • 张凤杰 ,
  • 崔路露 ,
  • 叶小侠 ,
  • 于佳俊 ,
  • 张晓蒙 ,
  • 杨武 ,
  • 王岸 ,
  • 薛洁
展开
  • 1(新疆农业大学 食品科学与药学学院,新疆 乌鲁木齐,830052)
    2(中国食品发酵工业研究院有限公司,北京,100015)
    3(房县黄酒产业发展中心,湖北 十堰,442100)
第一作者:硕士研究生(薛洁教授级高级工程师为通信作者,E-mail:lxxuejie@126.com)

收稿日期: 2025-04-04

  修回日期: 2025-07-23

  网络出版日期: 2026-04-10

基金资助

国家重点研发计划项目(2022YFD2101202)

收起

Research on freshness prediction model of Fangxian Huangjiu based on machine learning

  • HE Wenjie ,
  • ZHANG Fengjie ,
  • CUI Lulu ,
  • YE Xiaoxia ,
  • YU Jiajun ,
  • ZHANG Xiaomeng ,
  • YANG Wu ,
  • WANG An ,
  • XUE Jie
Expand
  • 1(School of Food Science and Pharmacy, Xinjiang Agricultural University, Urumqi 830052, China)
    2(China National Research Institute of Food & Fermentation industries Co.Ltd., Beijing 100015, China)
    3(Fangxian Huangjiu Industry Development Center, Shiyan 442100, China)

Received date: 2025-04-04

  Revised date: 2025-07-23

  Online published: 2026-04-10

Fold

摘要

该研究聚焦于房县黄酒的“鲜活”品质特征,利用机器学习算法构建了黄酒新鲜度预测模型。通过对76个样品的常规理化指标、氨基酸及挥发性风味物质的测定,并结合专家感官评分,筛选出L*值、b*值、电导率、非糖固形物和DPPH自由基清除率等关键变量作为模型输入。在随机森林、支持向量机回归和Lasso回归的对比中,随机森林模型表现最优,其决定系数(R2)达到0.89以上。特征重要性分析和SHAP(shapley additive explanations)解释方法表明,糠醛、L*值、b*值、硫代巴比妥酸值和乳酸乙酯对新鲜度影响显著,其中硫代巴比妥酸值、糠醛、b*值和L*值与新鲜度呈非线性关联。模型验证实验显示,随机抽取的10个未知样品的预测值与实测值偏差率为-6.30%~5.09%,验证了模型的可靠性。该研究为房县黄酒品质评价提供了客观量化的解决方案,有效解决了传统感官评价主观性强、稳定性差、难以量化以及对人员专业性要求高的问题,为其“鲜活”品质控制提供了重要参考依据。

关键词： 房县黄酒; 新鲜度; 随机森林; 支持向量机回归; Lasso回归

本文引用格式

何文杰 , 张凤杰 , 崔路露 , 叶小侠 , 于佳俊 , 张晓蒙 , 杨武 , 王岸 , 薛洁 . 基于机器学习的房县黄酒新鲜度预测模型研究[J]. 食品与发酵工业, 2026 , 52(6) : 136 -145 . DOI: 10.13995/j.cnki.11-1802/ts.042930

Abstract

This study investigated the characteristic “freshness” quality of Fangxian Huangjiu (a traditional Chinese rice wine) by developing a freshness prediction model using machine learning algorithms.Through measurements of conventional physicochemical parameters, amino acids, and volatile flavor compounds from 76 samples, alongside expert sensory evaluations, key input variables, including L* value, b*value, electrical conductivity, non-sugar solids, and DPPH free radical scavenging rate, were identified.Among comparative analyses of random forest (RF), support vector machine regression, and Lasso regression algorithms, the RF model demonstrated optimal performance with a coefficient of determination (R2) exceeding 0.89.Feature importance analysis and SHAP interpretation identified furfural, L*value, b* value, thiobarbituric acid (TBA) value, and ethyl lactate as the strongest predictors of freshness, with TBA value, furfural, b* value, and L* value exhibiting nonlinear relationships with freshness outcomes.Validation experiments on 10 randomly selected unknown samples showed deviation rates between predicted and measured values ranging from -6.30% to 5.09%, confirming model reliability.This research establishes an objective and quantifiable methodology for assessing Fangxian Huangjiu quality, effectively overcoming limitations associated with traditional sensory evaluation, such as subjectivity, instability, lack of quantifiability, and high dependency on professional expertise, thus providing critical reference standards for controlling its distinctive “freshness” quality.

Key words： Fangxian Huangjiu; freshness; random forest; support vector machine regression; Lasso regression

参考文献

[1] 毛健. 国酒:黄酒[M].北京:化学工业出版社, 2022:1-16.
MAO J.Huangjiu:Chinese National Alcohols[M].Beijing:Chemical Industry Press, 2022:1-16.
[2] 陈程. 房县黄酒[M].武汉:长江出版社, 2019.
CHEN C.Fangxian Huangjiu[M].Wuhan:Changjiang Press, 2019.
[3] 常堃, 陈祥, 李评, 等.基于SWOT分析的房县黄酒产业发展策略[J].中国酿造, 2025, 44(2):299-302.
CHANG K, CHEN X, LI P, et al.Development strategy of Fangxian Huangjiu industry based on SWOT analysis[J].China Brewing, 2025, 44(2):299-302.
[4] 刘慧杰, 王娟, 王忠朝, 等.全二维气相色谱-质谱联用技术解析房县黄酒的挥发性成分[J].食品与发酵工业, 2022, 48(17):249-255.
LIU H J, WANG J, WANG Z C, et al.Characterization of volatile compounds in Fangxian Huangjiu by two-dimensional gas chromatography-mass spectrometry[J].Food and Fermentation Industries, 2022, 48(17):249-255.
[5] GAO X L, FENG T, SHENG M J, et al.Characterization of the aroma-active compounds in black soybean sauce, a distinctive soy sauce[J].Food Chemistry, 2021, 364:130334.
[6] RICO-YUSTE A, GONZÁLEZ-VALLEJO V, BENITO-PEÑA E, et al.Furfural determination with disposable polymer films and smartphone-based colorimetry for beer freshness assessment[J].Analytical Chemistry, 2016, 88(7):3959-3966.
[7] 王楠, 张立福, 邓楚博, 等.光谱分析啤酒新鲜度检测方法[J].光谱学与光谱分析, 2020, 40(7):2273-2277.
WANG N, ZHANG L F, DENG C B, et al.Beer freshness detection method based on spectral analysis technology[J].Spectroscopy and Spectral Analysis, 2020, 40(7):2273-2277.
[8] 李鑫星, 董保平, 杨铭松, 等.基于SVM核机器学习的三文鱼新鲜度检测系统[J].农业机械学报, 2019, 50(5):376-384.
LI X X, DONG B P, YANG M S, et al.Detection system of salmon freshness based on SVM kernel-based machine learning[J].Transactions of the Chinese Society for Agricultural Machinery, 2019, 50(5):376-384.
[9] 蒋诗泉, 刘中侠, 蒋诗平, 等.随机森林算法在红葡萄酒质量评价指标体系选择中的应用[J].食品工业科技, 2014, 35(7):264-267.
JIANG S Q, LIU Z X, JIANG S P, et al.Application of random forest algorithm on selecting evaluation index system of the quality of red wine[J].Science and Technology of Food Industry, 2014, 35(7):264-267.
[10] RIBEIRO M N, CARVALHO I A, FONSECA G A, et al.Quality control of fresh strawberries by a random forest model[J].Journal of the Science of Food and Agriculture, 2021, 101(11):4514-4522.
[11] ASTRAY G, MEJUTO J C, MARTÍNEZ-MARTÍNEZ V, et al.Prediction models to control aging time in red wine[J].Molecules, 2019, 24(5):826.
[12] JU S, LIM H, MA J W, et al.Optimal county-level crop yield prediction using MODIS-based variables and weather data:A comparative study on machine learning models[J].Agricultural and Forest Meteorology, 2021, 307:108530.
[13] 刘竹青, 于梅.蛹虫草甜米酒的酿造工艺研究[J].中国酿造, 2017, 36(9):177-181.
LIU Z Q, YU M.Research on brewing technology of Cordyceps militaris sweet rice wine[J].China Brewing, 2017, 36(9):177-181.
[14] 朱正军, 万端极, 卢扬成, 等.传统房县黄酒酿造工艺研究[J].中国酿造, 2018, 37(2):95-99.
ZHU Z J, WAN D J, LU Y C.et al.Optimization of traditional brewing process of Fangxian rice wine[J].China Brewing, 2018, 37(2):95-99.
[15] 李运奎, 韩富亮, 张予林, 等.基于CIELAB色空间的红葡萄酒颜色直观表征[J].农业机械学报, 2017, 48(6):296-301.
LI Y K, HAN F L, ZHANG Y L, et al.Visualization for representation of red wine color based on CIELAB color space[J].Transactions of the Chinese Society for Agricultural Machinery, 2017, 48(6):296-301.
[16] 朱一松, 赵光鳌, 帅桂兰, 等.超滤法生产的纯生黄酒非生物稳定性的研究[J].食品与发酵工业, 2005(2):26-29.
ZHU Y S, ZHAO G A, SHUAI G L, et al.Study on the non-biological stability of draft rice wine by ultrafiltration[J].Food and Fermentation Industries, 2005(2):26-29.
[17] 肖侠. 啤酒表面张力与泡沫持久性关系的研究[D].兰州:甘肃农业大学, 2013.
XIAO X.Study on the relationship of surface tension and foam stability of beer[D].Lanzhou:Gansu Agricultural University, 2013.
[18] 严敏, 李崎, 顾国贤.一个评价啤酒老化程度的新指标──DPPH清除率[J].食品科技, 2006(3):91-93;100.
YAN M, LI Q, GU G X.A new index to estimate the extent of beer staling——the DPPH-scavenging ratio[J].Food Science and Technology, 2006(3):91-93;100.
[19] 于佳俊, 王德良, 秦谦, 等.强制老化过程中啤酒新鲜度、TBA值以及9种代表性老化物质的相关性研究[J].酿酒科技, 2014(8):1-5.
YU J J, WANG D L, QIN Q, et al.Study on the correlations among beer freshness, TBA values and 9 typical aging compounds during forced aging process[J].Liquor-Making Science & Technology, 2014(8):1-5.
[20] 沈棚, 薛红玮.陈酿时间对黄酒中挥发性风味物质和氨基酸含量的影响研究[J].中国酿造, 2023, 42(1):142-146.
SHEN P, XUE H W.Effect of aging time on volatile flavor substances and amino acids contents in Huangjiu[J].China Brewing, 2023, 42(1):142-146.
[21] NOGUEIRA A L, MUNITA C S.Quantitative methods of standardization in cluster analysis:Finding groups in data[J].Journal of Radioanalytical and Nuclear Chemistry, 2020, 325(3):1-6.
[22] UTKIN V L, KOVALEV M S, COOLEN F P A.Imprecise weighted extensions of random forests for classification and regression[J].Applied Soft Computing Journal, 2020, 92:106324.
[23] WEN C L, HUAI T T, ZHANG Q H, et al.A new rotation forest ensemble algorithm[J].International Journal of Machine Learning and Cybernetics, 2022, 13(11):3569-3576.
[24] GENUER R, POGGI J M, TULEAU-MALOT C.Variable selection using random forests[J].Pattern Recognition Letters, 2010, 31(14):2225-2236.
[25] ADHIKARI S, LECCI F, BECKER J T, et al.High-dimensional longitudinal classification with the multinomial fused lasso[J].Statistics in Medicine, 2019, 38(12):2184-2205.
[26] 陈镇, 刘润, 罗征, 等.基于机器学习的珠三角秋季臭氧浓度预测[J].环境科学, 2024, 45(1):1-7.
CHEN Z, LIU R, LUO Z, et al.Prediction of autumn ozone concentration in the Pearl River Delta based on machine learning[J].Environmental Science, 2024, 45(1):1-7.
[27] 张鹏泽, 叶丽, 姚军强, 等.基于机器学习的河流铈异常预测模型[J].环境科学学报, 2025, 45(5):337-346.
ZHANG P Z, YE L, YAO J Q, et al.Construction and analysis of machine learning based cerium anomaly prediction model for river[J].Acta Scientiae Circumstantiae, 2025, 45(5):337-346.
[28] ZHAO B, ZHU W X, HAO S F, et al.Prediction heavy metals accumulation risk in rice using machine learning and mapping pollution risk[J].Journal of Hazardous Materials, 2023, 448:130879.
[29] 徐建芬, 张凤杰.黄酒陈酿过程中香气物质、味觉特性和表面张力变化[J].中国酿造, 2018, 37(1):41-44.
XU J F, ZHANG F J.Changes of aroma substances, taste characteristics and surface tension during the aging process of Chinese rice wine[J].China Brewing, 2018, 37(1):41-44.
[30] 王岸娜, 吴立根.贮存时间对黄酒成分的影响[C].2010年国际细胞生物学、生物学、生物工程会议.Piscataway:IEEE, 2011.
WANG A N, WU L G.The influence of storage time on the composition of yellow wine[C].2010 International Conference on Cellular, Molecular Biology, Biophysics and Bioengineering.Piscataway:IEEE, 2011.
[31] 朱长满. 煎酒对客家黄酒品质及其抗氧化活性的影响[D].广州:仲恺农业工程学院, 2019.
ZHU C M.Effects of decoction process on the quality and antioxidant activity of hakka rice wine[D].Guangzhou:Zhongkai University of Agriculture and Engineering, 2019.
[32] ZHAO C J, SCHIEBER A, GÄNZLE M G.Formation of taste-active amino acids, amino acid derivatives and peptides in food fermentations-A review[J].Food Research International, 2016, 89(P1):39-47.
[33] SONG G S, JIANG N L, ZHAO L W, et al.Analysis of aroma characteristics of the aged Shaoxing Huangjiu by a combination of multiple instruments and machine learning[J].LWT, 2024, 214:117165.
[34] 蒋洋, 张翠英, 李于, 等.酒类风味物质对人体乙醇代谢影响的研究进展[J].食品科学, 2021, 42(15):242-250.
JIANG Y, ZHANG C Y, LI Y, et al.Effect of flavor substances in alcoholic beverages on ethanol metabolism in human body:A review[J].Food Science, 2021, 42(15):242-250.
[35] LI M, ZHAN P, WANG P, et al.Characterization of aroma-active compounds changes of Xiecun Huangjius with different aging years based on odor activity values and multivariate analysis[J].Food Chemistry, 2022, 405(PA):134809.
[36] WANG J, YU Y G, GAO X L, et al.Succession patterns of aroma components during brewing process of broomcorn millet (Panicum miliaceum L.) Huangjiu[J].Food Research International, 2022, 154:110982.
[37] 郑翠银, 龚丽婷, 黄志清, 等.甜型红曲黄酒中关键挥发性香气成分分析[J].中国食品学报, 2014, 14(5):209-217.
ZHENG C Y, GONG L T, HUANG Z Q, et al.Analysis of the key volatile aroma compounds in sweet Hongqu glutious rice wine[J].Journal of Chinese Institute of Food Science and Technology, 2014, 14(5):209-217.
Options
文章导航

/

〈 〉

技术支持：北京玛格泰克科技发展有限公司