CN113935573A - Intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with delicate fragrance - Google Patents
Intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with delicate fragrance Download PDFInfo
- Publication number
- CN113935573A CN113935573A CN202111038566.XA CN202111038566A CN113935573A CN 113935573 A CN113935573 A CN 113935573A CN 202111038566 A CN202111038566 A CN 202111038566A CN 113935573 A CN113935573 A CN 113935573A
- Authority
- CN
- China
- Prior art keywords
- liquor
- fen
- pca
- sensory
- fenjiu
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000001953 sensory effect Effects 0.000 title claims abstract description 68
- 238000011156 evaluation Methods 0.000 title claims abstract description 39
- 239000003205 fragrance Substances 0.000 title description 4
- 238000000513 principal component analysis Methods 0.000 claims abstract description 62
- 239000000796 flavoring agent Substances 0.000 claims abstract description 38
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000004458 analytical method Methods 0.000 claims abstract description 23
- 238000005516 engineering process Methods 0.000 claims abstract description 23
- 238000005481 NMR spectroscopy Methods 0.000 claims abstract description 20
- 238000007619 statistical method Methods 0.000 claims abstract description 14
- 238000005160 1H NMR spectroscopy Methods 0.000 claims abstract description 13
- 238000001228 spectrum Methods 0.000 claims abstract description 11
- 238000012795 verification Methods 0.000 claims abstract description 11
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 125000004435 hydrogen atom Chemical group [H]* 0.000 claims abstract description 8
- UFHFLCQGNIYNRP-UHFFFAOYSA-N Hydrogen Chemical compound [H][H] UFHFLCQGNIYNRP-UHFFFAOYSA-N 0.000 claims abstract description 3
- 239000001257 hydrogen Substances 0.000 claims abstract description 3
- 229910052739 hydrogen Inorganic materials 0.000 claims abstract description 3
- 238000002790 cross-validation Methods 0.000 claims description 19
- 239000000126 substance Substances 0.000 claims description 12
- 230000010354 integration Effects 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 7
- 238000010200 validation analysis Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000000655 nuclear magnetic resonance spectrum Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000009825 accumulation Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims 1
- 238000007796 conventional method Methods 0.000 abstract description 2
- 235000014101 wine Nutrition 0.000 description 51
- 239000000523 sample Substances 0.000 description 14
- 238000011160 research Methods 0.000 description 13
- HEMHJVSKTPXQMS-UHFFFAOYSA-M Sodium hydroxide Chemical compound [OH-].[Na+] HEMHJVSKTPXQMS-UHFFFAOYSA-M 0.000 description 9
- 235000013305 food Nutrition 0.000 description 8
- 235000019634 flavors Nutrition 0.000 description 7
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000005033 Fourier transform infrared spectroscopy Methods 0.000 description 6
- 239000000243 solution Substances 0.000 description 6
- 235000015096 spirit Nutrition 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 230000004069 differentiation Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 4
- XLYOFNOQVPJJNP-ZSJDYOACSA-N Heavy water Chemical compound [2H]O[2H] XLYOFNOQVPJJNP-ZSJDYOACSA-N 0.000 description 4
- 238000004497 NIR spectroscopy Methods 0.000 description 4
- NBIIXXVUZAFLBC-UHFFFAOYSA-N Phosphoric acid Chemical compound OP(O)(O)=O NBIIXXVUZAFLBC-UHFFFAOYSA-N 0.000 description 4
- 238000004611 spectroscopical analysis Methods 0.000 description 4
- 241000196324 Embryophyta Species 0.000 description 3
- VEXZGXHMUGYJMC-UHFFFAOYSA-N Hydrochloric acid Chemical compound Cl VEXZGXHMUGYJMC-UHFFFAOYSA-N 0.000 description 3
- PXIPVTKHYLBLMZ-UHFFFAOYSA-N Sodium azide Chemical compound [Na+].[N-]=[N+]=[N-] PXIPVTKHYLBLMZ-UHFFFAOYSA-N 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 239000007853 buffer solution Substances 0.000 description 3
- 150000001875 compounds Chemical class 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 229910000402 monopotassium phosphate Inorganic materials 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 238000004451 qualitative analysis Methods 0.000 description 3
- 238000004445 quantitative analysis Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 235000019640 taste Nutrition 0.000 description 3
- DGAQECJNVWCQMB-PUAWFVPOSA-M Ilexoside XXIX Chemical compound C[C@@H]1CC[C@@]2(CC[C@@]3(C(=CC[C@H]4[C@]3(CC[C@@H]5[C@@]4(CC[C@@H](C5(C)C)OS(=O)(=O)[O-])C)C)[C@@H]2[C@]1(C)O)C)C(=O)O[C@H]6[C@@H]([C@H]([C@@H]([C@H](O6)CO)O)O)O.[Na+] DGAQECJNVWCQMB-PUAWFVPOSA-M 0.000 description 2
- 239000007836 KH2PO4 Substances 0.000 description 2
- XLOMVQKBTHCTTD-UHFFFAOYSA-N Zinc monoxide Chemical compound [Zn]=O XLOMVQKBTHCTTD-UHFFFAOYSA-N 0.000 description 2
- 229910000147 aluminium phosphate Inorganic materials 0.000 description 2
- 238000005102 attenuated total reflection Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000007621 cluster analysis Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000001506 fluorescence spectroscopy Methods 0.000 description 2
- 238000002290 gas chromatography-mass spectrometry Methods 0.000 description 2
- 150000002500 ions Chemical class 0.000 description 2
- 238000004949 mass spectrometry Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000000491 multivariate analysis Methods 0.000 description 2
- 239000005416 organic matter Substances 0.000 description 2
- 239000012071 phase Substances 0.000 description 2
- PJNZPQUBCPKICU-UHFFFAOYSA-N phosphoric acid;potassium Chemical compound [K].OP(O)(O)=O PJNZPQUBCPKICU-UHFFFAOYSA-N 0.000 description 2
- 238000012628 principal component regression Methods 0.000 description 2
- 238000000425 proton nuclear magnetic resonance spectrum Methods 0.000 description 2
- 150000003839 salts Chemical class 0.000 description 2
- 229910052708 sodium Inorganic materials 0.000 description 2
- 239000011734 sodium Substances 0.000 description 2
- 230000008685 targeting Effects 0.000 description 2
- 239000012855 volatile organic compound Substances 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000005303 weighing Methods 0.000 description 2
- 235000014692 zinc oxide Nutrition 0.000 description 2
- 239000011787 zinc oxide Substances 0.000 description 2
- 240000004246 Agave americana Species 0.000 description 1
- 229910019142 PO4 Inorganic materials 0.000 description 1
- 235000008331 Pinus X rigitaeda Nutrition 0.000 description 1
- 235000011613 Pinus brutia Nutrition 0.000 description 1
- 241000018646 Pinus brutia Species 0.000 description 1
- XBDQKXXYIPTUBI-UHFFFAOYSA-N Propionic acid Chemical class CCC(O)=O XBDQKXXYIPTUBI-UHFFFAOYSA-N 0.000 description 1
- 239000005708 Sodium hypochlorite Substances 0.000 description 1
- 238000004224 UV/Vis absorption spectrophotometry Methods 0.000 description 1
- 235000009754 Vitis X bourquina Nutrition 0.000 description 1
- 235000012333 Vitis X labruscana Nutrition 0.000 description 1
- 240000006365 Vitis vinifera Species 0.000 description 1
- 235000014787 Vitis vinifera Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 235000001014 amino acid Nutrition 0.000 description 1
- 150000001413 amino acids Chemical class 0.000 description 1
- 229930002877 anthocyanin Natural products 0.000 description 1
- 235000010208 anthocyanin Nutrition 0.000 description 1
- 239000004410 anthocyanin Substances 0.000 description 1
- 150000004636 anthocyanins Chemical class 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 125000003118 aryl group Chemical group 0.000 description 1
- 235000019996 baijiu Nutrition 0.000 description 1
- 235000013532 brandy Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 238000012569 chemometric method Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000005686 electrostatic field Effects 0.000 description 1
- -1 ester compounds Chemical class 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000855 fermentation Methods 0.000 description 1
- 230000004151 fermentation Effects 0.000 description 1
- 238000001319 headspace solid-phase micro-extraction Methods 0.000 description 1
- 238000004896 high resolution mass spectrometry Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000001819 mass spectrum Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000019796 monopotassium phosphate Nutrition 0.000 description 1
- 238000001208 nuclear magnetic resonance pulse sequence Methods 0.000 description 1
- 150000007524 organic acids Chemical class 0.000 description 1
- 235000005985 organic acids Nutrition 0.000 description 1
- 238000010239 partial least squares discriminant analysis Methods 0.000 description 1
- 238000010238 partial least squares regression Methods 0.000 description 1
- 150000002989 phenols Chemical class 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000004513 sizing Methods 0.000 description 1
- SUKJFIGYRHOWBL-UHFFFAOYSA-N sodium hypochlorite Chemical compound [Na+].Cl[O-] SUKJFIGYRHOWBL-UHFFFAOYSA-N 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000002904 solvent Substances 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012916 structural analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 235000019605 sweet taste sensations Nutrition 0.000 description 1
- 235000013529 tequila Nutrition 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000001269 time-of-flight mass spectrometry Methods 0.000 description 1
- 125000000026 trimethylsilyl group Chemical group [H]C([H])([H])[Si]([*])(C([H])([H])[H])C([H])([H])[H] 0.000 description 1
- 238000001195 ultra high performance liquid chromatography Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
- 235000015041 whisky Nutrition 0.000 description 1
- 238000004149 wine analysis Methods 0.000 description 1
- 239000010151 yanghe Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N24/00—Investigating or analyzing materials by the use of nuclear magnetic resonance, electron paramagnetic resonance or other spin effects
- G01N24/08—Investigating or analyzing materials by the use of nuclear magnetic resonance, electron paramagnetic resonance or other spin effects by using nuclear magnetic resonance
- G01N24/087—Structure determination of a chemical compound, e.g. of a biomolecule such as a protein
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2132—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on discrimination criteria, e.g. discriminant analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Economics (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- High Energy & Nuclear Physics (AREA)
- Molecular Biology (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Crystallography & Structural Chemistry (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides an intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor, belonging to the technical field of liquor quality analysis, and the method comprises the following steps: (1) using one-dimensional nuclear magnetic resonance hydrogen spectrum (1H NMR) technology to determine the characteristic fingerprint information of hydrogen atoms of organic matters in the fen-liquor base liquor; (2)1preprocessing an H NMR spectrogram; (3) establishing a fen-flavor Fenjiu sensory grade PCA/LDA model by adopting a multivariate statistical method combining unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA); (4) and verifying and evaluating the effectiveness of the PCA/LDA model by adopting an internal leave-one cross verification method and an external repeated double-random cross verification method. The invention establishes faint scent through the characteristic fingerprint of hydrogen atoms of organic matters of the faint scent Fenjiu base liquor and combining a multivariate statistical meansThe sensory intelligent digital grading technology of Fenjiu base liquor breaks through the technical limitation that effective intelligent and digital grading evaluation on sensory quality grades of white liquor base liquor is difficult to carry out by adopting conventional methods and means in the past.
Description
Technical Field
The invention belongs to the technical field of white spirit quality analysis, and particularly relates to an intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor.
Background
Fen-flavor liquor is popular with consumers for a long time due to its long history and unique flavor of soft taste, sweet taste and pure fragrance, and the taste, flavor and price of the liquor with different grades are very different [ Hoffodian. Shanxi fen-flavor liquor characteristic fragrance component discovery and sensory and intelligent instrument comprehensive analysis [ D ] in Genzhong: Shanxi agriculture university, 2018 ]. The quality and grade of the white spirit base liquor determine the quality and grade of finished liquor to a certain extent, the base liquor with different quality grades can blend finished liquor with different styles, different tastes and different grades, and the accurate identification of the quality grade of the white spirit base liquor is an important basis for graded storage and blending of the white spirit [ Sunzong, Zhouxuan, Wujianfeng, and the like. At present, the quality grade of the white spirit base spirit is still mainly identified by depending on the detection and sensory evaluation of physicochemical indexes by white spirit enterprises, but certain individual difference, instability and limitation exist in artificial sensory evaluation, the evaluation standard caused by subjective factors is not uniform, so that the difference among products in different grades is difficult to quantify and standardize, time and labor are consumed, batch completion is also impossible, the market differentiation requirements are difficult to meet, and the method is a difficult problem to solve in the field of intelligent manufacturing of the current Chinese white spirit. Therefore, the development of an intelligent and digital grading evaluation technology for the sensory quality of Chinese liquor base wine is particularly necessary.
In the aspect of liquor quality grade identification, Zhang Kexin et al (Zhang Kexin, Lidong and winter, action of base liquor evaluation in liquor production process [ J ]. modern salt chemical industry, 2018.) propose that sensory evaluation is an important basis for determining base liquor quality grade and selecting high-grade liquor products, and is a prerequisite for product sizing. Yangdi adopts Near Infrared spectroscopy (NIR) technology to research the content change rule of flavor components in the white spirit base wine, and combines a chemometrics means to realize the rapid grading of the white spirit base wine grade [ Yangdi. establishment of a white spirit base wine analysis model based on the Near Infrared spectroscopy [ D ]. Henan science and technology university, 2016 ]. Sunzong et al use Fourier Transform Infrared spectroscopy (FTIR) and Attenuated Total Reflection (ATR) techniques to perform rapid qualitative and quantitative Analysis on different grades of liquor base and its main ester compounds, construct Linear Discriminant Analysis (LDA) and Back Propagation Artificial Neural Network (BP-ANN) chemometric models, and realize 100% identification of different grades of liquor base [ Sunzong, Xinxin, wavelet, etc.. Fourier Transform Infrared spectroscopy combines with chemometric methods to perform rapid qualitative and quantitative Analysis [ J ] spectroscopy and spectroscopic Analysis [ 2017 ]. Chengdong et al detected ion abundance values of fen-flavor type Yanghe Daqu liquor samples of different grades and Niuba mountain liquor mass-to-charge ratio (m/z) within 55-191 by adopting a headspace solid phase microextraction mass spectrometry (HS-SPME-MS) technology, screened out important characteristic ions by combining Partial Least Squares Discriminant Analysis (PLS-DA), Stepwise Linear Discriminant Analysis (SLDA), Partial Least Squares Regression Analysis (PLS-Least Regression) and Principal Component Regression Analysis (PCR), constructed ANN and Support Vector Machine (SVM) models, realized mass spectrometry [ Pingtheng, Normal, Xuyan, et al ] based on the liquor grade and chemical metrology and flavor type [ J ] and industrial grade identification [ 3. Baijiu ] fermentation grade identification ] of liquor grade greater than 86%, fang wen lai, xu yan, etc. fen-flavor liquor grade discrimination based on mass spectrum and support vector machine [ J ] food industry science and technology, 2014 ]. Zhongying et al propose a fuzzy comprehensive evaluation method to describe flavor differences of the fragrant aroma type Xiangquan base liquor, and determine the quality grade of the fragrant aroma type Xiangquan base liquor according to the membership degree calculated by using a membership function [ Zhongying, Mo xian is cheap, and di-zhang. application of fuzzy comprehensive evaluation in Xiangquan base liquor quality sensory evaluation [ J ]. brewing, 2005 ]. Congratulatory et al established GC fingerprint spectra of 3 types of strong aromatic 1-year old liquor base liquors with different grades, and identified liquor base liquors with different sensory quality grades by adopting an included angle cosine and correlation coefficient method [ congratulation, Zhangyi, Zhaojin pine, similarity research of liquor base liquors with different sensory grades [ J ]. Chinese brewing, 2011 ]. Plum source wine adopts a GC technology to carry out qualitative detection and analysis on volatile and semi-volatile components in the fragrant odor type drunkard wine base wine and finished wine, and establishes a GC fingerprint for identifying and evaluating the quality grade of the drunkard wine base wine and the quality of the finished wine [ study of fragrant odor type drunkard wine flavor substance fingerprint [ D ]. Hunan university, 2011 ]. Zhou Xuan adopts HS-SPME-MS technology to carry out qualitative and quantitative analysis on volatile components in base liquor of Luzhou-flavor liquor with different quality grades, and establishes a method for quickly grading the quality of the base liquor of Luzhou-flavor liquor based on a volatile component database of the base liquor of liquor [ Zhou Xuan. base liquor of Luzhou-flavor liquor volatile component analysis and grade recognition research [ D ]. Jiangsu university, 2019 ]. Molley and the like adopt an ultra-high performance liquid chromatography-quadrupole/electrostatic field orbitrap high-resolution mass spectrometry method to establish a flavor substance database of the fragrant scent type white spirit, directly identify and identify trace components in 73 white spirits, and realize good differentiation of different grades of the fragrant scent type white spirit by combining a principal component analysis method [ Molley, Qianjiu, Lishaohui, and the like ] research on grade identification of the fragrant scent type white spirit [ J ] food safety quality detection statement, 2018 ].
Internationally, no research report on the sensory quality grading of the fen-flavor liquor is found, but some research reports on the application research on the quality and grade differentiation of the spirits and the wines exist, and a large amount of research data and experience references are provided for the invention. Palma et al predicted the age of spirits brandy and wine using FTIR techniques, and the resultsWine age was found to have a high correlation with FTIR data [ palm M, Barroso CG. application of FT-IR spectroscopy to the characterization and classification of wires, brandes and other distilled drugs [ J].Talanta,2002.]. Barbosa-Garci et al used UV-visible spectroscopy to identify different brands of New Agave wine [ Barbosa-Garcia, G.Ramos-Ort i z, Maldonao J L, et al.UV-vis absorption spectroscopy and multivariates analysis as a method to discard tequila [ J].SpectrochimicaActa Part A Molecular&Biomolecular Spectroscopy,2007.]. The types of four liquors were correctly classified and adulterated by Pontes et al using NIR in combination with chemometrics [ M. sub. Jos. Coelho Pontes, Galv O R K H, M. sub. C. sarugino Ara jo, et al. the scientific projects for specific variable selection in classification schemes [ J. J. ]].Chemometrics&Intelligent Laboratory Systems,2005]. Sadecka et al used fluorescence spectroscopy and multivariate data Analysis to distinguish between brands and distilled spirits of different types, PCA and Cluster Analysis (CA) showed that spectral data with an excitation wavelength of 350nm and an emission wavelength of 360-650nm gave the best results in distinguishing between brands and distilled spirits [ Sadecka J, Tothova J, Majek P].Food Chemistry,2009.]. Dufour et al detected and analyzed wine samples of different varieties, grades and years by fluorescence spectroscopy to obtain fingerprint information of different wine samples for distinguishing different varieties and grades of wines [ Dufour, Ledit A, Laguet A, et al].AnalyticaChimicaActa,2006.]. Cardeal et al used comprehensive two-dimensional gas chromatography-mass spectrometry (GCxGC/TOFMS) to analyze and compare volatile organic compounds of Kasha liquor and spirits in Brazil to determine the difference and similarity of spirits [ Cardeal ZL (Univ Fed Minas Gerais), Marriott PJ.compressive two-dimensional gas chromatography-mass spectrometry analysis and compliance of volatile organic compounds in Brazilian scales and selected ones [ J].Food Science and Technology Abstracts,2009.]. Collins et al used an ultra-high performance liquid chromatography time-of-flight mass spectrometer in combination to analyze 63 US whisky nonvolatile fractions and discriminate the differences between samples [ Collins TS (Univ Calif Davis), Zweigenbaum J, Ebeler SE].Food Chemistry,2014.]. German Bruker establishes wine non-targeting1H NMR fingerprint spectrum technology realizes authenticity identification of more than 90% of wine year, variety and production area [ Godelmann R, Fang F, Humpfer E, et al1H NMR spectroscopy combined with multivariate statistical analysis.Differentiation of Important Parameters:Grape Variety,Geographical Origin,Year of Vintage[J].Journal of Agricultural&Food Chemistry,2013.]。1The main advantages of the H NMR technology are that the fingerprint data contain a large amount of structural information of unknown compounds, which is beneficial to the structural analysis of the unknown compounds, and the signals can provide high-flux data of volatile compounds, phenols, organic acids, anthocyanins, amino acids and the like, compared with the traditional analysis means,1h NMR detection can reflect all the characteristics of the white spirit ingredients. On the basis of this, the method is suitable for the production,1the H NMR technology is expected to provide a new idea and method for the level digital grading of the sensory quality of the white spirit.
Disclosure of Invention
The invention takes fen-liquor base liquor of fen-liquor with faint scent as a research object and adopts1H NMR technology is used for measuring characteristic fingerprint information of hydrogen atoms of organic matters in fen-flavor Fenjiu base liquor, data mining is carried out by combining a multivariate statistical method, and a fen-flavor Fenjiu base liquor sensory intelligent digital grading model is established, so that the technical limitation that effective intelligent and digital grading evaluation is difficult to carry out on sensory quality grades of the fen-flavor Fenjiu base liquor by adopting conventional methods and means in the past is broken through.
Specifically, the invention provides an intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with faint scent, which comprises the following steps of1H NMR technology for determining organic matters in fen-flavor liquor baseThe method comprises the steps of carrying out atlas preprocessing and segmentation integration on physical hydrogen atom characteristic fingerprint information, carrying out standardization and centralization preprocessing on data after the segmentation integration, establishing a fen-flavor liquor sensory grade PCA/LDA model by adopting a multivariate statistical method combining unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA), verifying the PCA/LDA model by adopting an internal one-leave-one cross verification method and an external repeated double-random cross verification method, and evaluating the correct classification capability and stability of the PCA/LDA model by the average value, standard deviation, median and absolute standard deviation of median of the correct classification rate of the forecast grade and the actual sensory quality grade of the fen-flavor liquor sensory grade PCA/LDA model.
The specific technical scheme of the invention is as follows:
the invention provides an intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor, which comprises the following steps:
(1) using one-dimensional nuclear magnetic resonance hydrogen spectrum (1H NMR) technology to determine the characteristic fingerprint information of hydrogen atoms of organic matters in the fen-liquor base liquor to obtain a spectrogram;
(2) performing spectrum pretreatment and segmentation integration on the spectrogram;
(3) carrying out data preprocessing on the data obtained by the segmented integration;
(4) establishing a fen-flavor Fenjiu sensory quality grade and a fen-flavor Fenjiu sensory grade PCA/LDA model of data obtained by the segmented integration in the step (3) by adopting a multivariate statistical method combining unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA);
(5) verifying the PCA/LDA model obtained in the step (4) by adopting an internal one-left cross verification method and an external repeated double-random cross verification method;
(6) evaluation of the PCA/LDA model described above: and evaluating the correct classification capability and stability of the PCA/LDA model through the average value, the standard deviation, the median and the absolute standard deviation of the median of the correct classification rates of the prediction grade and the actual sensory quality grade of the fen-liquor base liquor sensory grade PCA/LDA model.
Further, in the step (2), the spectrum preprocessing comprises Fourier transform, baseline correction, phase adjustment, chemical shift calibration and area segmentation integral correction.
Further, in step (3), the data preprocessing includes zero-mean normalization of the data.
Further, in the step (4), when the fen-flavor Fenjiu sensory grade PCA/LDA model is established, the fen-flavor Fenjiu base liquor is used1Taking segmented integral Bins data of the H NMR spectrum as an independent variable X matrix, and establishing an unsupervised PCA model;
selecting a principal component score matrix with the variable accumulation variance contribution rate of more than 95% in the PCA model as an input variable, selecting the sensory quality grade of the fen-liquor base liquor with the fen-flavor type established by a sensory evaluation team as a classification standard of the supervised LDA model, and establishing the supervised LDA model.
Further, in the step (5), the repeated double-random cross validation specifically includes selecting a random hierarchical sampling mode, randomly extracting 70% of data in an original data set as a training set for modeling, 30% of data as an external validation set for prediction, and repeatedly sampling 100 times to validate the supervised LDA model.
Further, in the step (6), the more the average value of the correct classification rate is close to 1, the more the median is close to 1, the smaller the standard deviation value is, the smaller the absolute standard deviation value of the median is, and the stronger the stability of the correct classification capability of the model is.
Furthermore, when the sensory quality grade of the fen-flavor Fenjiu base liquor is established by a sensory evaluation team, the quality grade of the Fenjiu is divided according to sensory scores mainly according to the enterprise standard of Fenjiu newly-produced liquor Q/XFJ 06.04-2019.
The invention has the following advantages:
the invention provides an intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with fen-flavor, aiming at the problems that individual difference and instability phenomena exist in manual sensory evaluation of liquor product quality grades, and quantification and standardization of difference among liquor products with different quality grades are urgently needed in the field of intelligent manufacturing of Chinese liquor at present so as to meet the differentiation requirements of markets, wherein a PCA/LDA model for the sensory quality of fen-liquor base liquor with fen-flavor is established on the basis of a hydrogen atom characteristic fingerprint of an organic matter in fen-liquor with fen-flavor by a multivariate statistical method combining unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA), and the model can realize that the average correct classification rate between a predicted quality grade and an actual quality grade is 92.2%, and has a good digitalized grading evaluation effect.
The invention firstly proposes that the non-target nuclear magnetic resonance organic matter hydrogen atom characteristic fingerprint technology is combined with a multivariate statistical analysis means to carry out intelligentized and digitized grading on the sensory quality of the fen-flavor liquor base, lays a good work foundation for constructing a software system for intelligentized and digitized graded evaluation of the sensory quality of the fen-flavor liquor base, is beneficial to improving the quality and the quality of the fen-liquor product, deepens the research and the application of modern instrument analysis technology in the field of intelligent sensory evaluation of liquor, promotes the progress of intelligent manufacture in the Chinese liquor industry, and helps the high-quality healthy sustainable development of the liquor industry.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate an embodiment of the invention and, together with the description, serve to explain the invention and not to limit the invention.
In the drawings:
FIG. 1 is a graph showing the scores of principal component 1 and principal component 2 of Fenjiu base liquor samples of different sensory grades in examples of the present invention.
FIG. 2 is a linear discrimination score chart of the PCA/LDA model for Fenjiu base wine samples of different sensory grades in the example of the present invention.
FIG. 3 is a flow chart of a PCA/LDA model verification and evaluation method in an embodiment of the present invention.
Detailed Description
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
Example 1An intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor.
1 materials and methods
1.1 materials
14 samples of the first-grade liquor of the white spirit base liquor, 59 high-grade liquors, 18 special-grade liquors and 92 samples in total are provided by Shanxi Xinghua cunfen winery GmbH.
1.2 instrumentation
Bruker Avance III HD 400MHz NMR spectrometer, Bruker 582-bit autosampler (Samplejet) and Bruker 5mm high throughput NMR tube (Bruker Biospin, Germany); pH meter (Mettler-Toledo, Switzerland).
1.3 reagents
Heavy water (D)2O, 99.9%, Sigma-Aldrich, china, shanghai); sodium 3- (trimethylsilyl) deuterated propionate (Sodium-3- (trimethylsilyl) propionate-2,2,3,3-d4, TSP, 98%, annolon (beijing) biotechnology limited); potassium dihydrogen phosphate (KH)2PO4, analytical grade, beijing chemical plant); sodium azide (super pure, NaN)3Beijing Bootoda technologies, Inc.); sodium hydroxide (NaOH, analytical grade, beijing chemical plant); hydrochloric acid (HCl, analytical pure, beijing chemical plant).
1.4 methods
1.4.1 buffer solution preparation
Accurately weighing 1.0g TSP and 0.13g NaN respectively3In a beaker, with D2Dissolving the sodium hypochlorite solution in a 10mL volumetric flask to obtain 100g/L TSP solution and 13g/L NaN3And (3) solution. Accurately weighing 8.0g KH2PO4In a beaker, 100mL of D2O was dissolved and transferred to a 200mL volumetric flask, to which 5mL of phosphoric acid, 2mL of the above TSP solution and 2mL of NaN were added3Solution, and 50mL of D2And O. After standing for 24 hours, the pH of the solution is measured, if it is pH>2.0, adding proper amount of phosphoric acid to adjust the pH value<2.0, then adding a proper amount of KH2PO4And adjusting the solid powder until the pH value is stabilized to 2.0, and finishing the preparation of the buffer solution.
1.4.2 sample preparation
Accurately sucking 300 μ L of buffer solution into a sample tube, adding 2700 μ L of Chinese liquor sample, mixing, and precisely adjusting sample with 3M NaOH and HClProduct pH to 4.0 (+ -0.02), accurately pipette 600. mu.L of sample mixture into a 5mm NMR tube for1H NMR measurement.
1.4.3 1H NMR measurement parameters
For experiments1H NMR at 400.13MHz with a PA BBI 400S 1H-BB-D-05Z probe; the detection temperature of the instrument is 300K; a Spectral Width (SW) of 20.5524 ppm; the number of sampling points (TD) is 65536 (64K); the spectral line broadening factor is 0.3 Hz; a Reception Gain (RG) of 16; the number of free induction decay scans (NS) per scan was 64; a relaxation delay (D1) of 4 s; sodium 3- (trimethylsilyl) deuteropropionate (TSP, δ ═ 0) was used as the zero point of the chemical shift. Pulse sequences of the Bruker standard (NOESYGPPS 1D) were used for characteristic peak signal suppression for water (δ ═ 4.8) and ethanol (δ ═ 1.18 and δ ═ 3.64). Before the sample is detected, the temperature is stabilized for 5 minutes.
1.4.4 1H NMR spectrum pretreatment
1H NMR signal deviation is limited by various factors such as the fluctuation of the instrument, the salt solubility, the temperature, the pH value and the like, and proper ones are selected1The pretreatment method of the H NMR sample spectrum can obviously reduce1H NMR spectrum peak drift, uneven base line, asymmetric peak shape, solvent peak interference and the like.
The invention adopts MestReNova 12.0 software (Mestrelab Research S.L., MestReNova (Mnova) NMR, USA) pair1The H NMR spectra were subjected to fourier transform, baseline correction, phase adjustment and calibration of chemical shifts. Based on R software, using CLUPA algorithm pair1Carrying out regional segmentation alignment correction on the chemical displacement offset of characteristic peaks of the H NMR spectrum1The H NMR spectrum is subdivided into a plurality of regions (Bins), the integral values of all points of each region adding up to represent the entire original spectrum in an abstract way, and the integral Bins of these regions are then used as input variables for statistical analysis modeling [ Esslinger S, Riedl J, Fauhl-Hassek C].Food Research International,2014.]。
1.4.5 data quality control
92 samples of Fenjiu base liquor1And (4) overlaying HNMR (human-nuclear magnetic resonance) spectra, and further checking whether the samples exist due to improper inhibition of main components of water and ethanol in the white spirit, system errors or other reasons1The HNMR atlas has abnormal baseline deviation and fluctuation, and each sample is repeatedly measured for 2 times, so that the stability, reliability and accuracy of data quality are ensured.
1.4.6 Fenjiu base wine sample sensory rating and grading
The sensory evaluation team is formed by six fenjiu professional tasters, and the sensory evaluation team is in turn matchedFenjiu base wineAnd (4) carrying out blind evaluation, carrying out sensory scoring, taking an average value, and recording related sensory comments. According to the enterprise standard of Fenjiu newly-produced wine Q/XFJ 06 06.04-2019, the newly-produced Fenjiu quality is divided into three grades according to sensory scores, A: first-class wine (61-64 min); b: top-grade wine (65-79 points); c: super wine (more than or equal to 80 points).
1.4.7 multivariate statistical analysis
Multivariate statistical analysis includes unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA). In order to facilitate the modeling of statistical analysis, three quality grades of newly produced fen wine are subjected to digital grading, and first-grade wine (A), superior-grade wine (B) and special-grade wine (C) respectively correspond to 1, 2 and 3 in sequence.
PCA as the first step of exploratory data dimensionality reduction determines the principal component matrix from the 95% variance in the interpretation dataset. The principal component score matrix can be used as an input variable of the LDA, and the principal components are mutually orthogonal, so that the problem of high collinearity among initial variables in the data set can be solved.
LDA can maximize the difference between dependent variables, minimize the sample distance within the group, and improve the effectiveness and analytical capability of the model [ Mardia K V, Kent J T, Bibby, J M.
PCA and LDA data analysis were both done by R software version 3.2.3 processing [ R Core team.R: A language and environment for statistical computing. https:// www.R-project. org. ], using R software data packets of mdatools [ Kucheryavskiy S.mdatools: multivariate data analysis for chemometrics. https:// CRAN.R-project. org/package ═ mdatools ] and mass [ Venabels W, Ripley B, et al model applied statistics with S [ M ].4 th. Berlin: Springer,2002 ]. All data were normalized by R software using a zero-mean normalization method (standard deviation 1, mean 0) prior to PCA and LDA analysis. Evaluation of the effectiveness of the PCA/LDA model was performed using a one-by-one cross-validation [ Stone M.Cross-validation and assessment of Statistical predictions [ J ]. Journal of the Royal Statistical Society,1974 ] and a repeated double random cross-validation [ Filzmoser P, Liebmann B, Varmuza K.repeated double cross-validation [ J ]. Journal of Chemometrics,2009 ] all done by R software.
2 detailed operational analysis and results
2.2.1 Principal Component Analysis (PCA)
Will be provided with1Carrying out data preprocessing operations such as peak alignment, segmented integration and the like on the H NMR spectrogram, and then sampling the fen-liquor base liquor with the faint scent type1The H NMR data are introduced into R data analysis software, PCA analysis is used for exploring the possibility of distinguishing samples of the Fenjiu base wine with different sensory grades, as shown in figure 1, the unsupervised grouping situation of the samples of the Fenjiu base wine with different sensory grades in a two-dimensional coordinate system consisting of PC1(Comp 1) and PC2(Comp 2) is preliminarily reflected in figure 1, and meanwhile, the cumulative contribution rate of the variance of the explained data of the first two main components of the PCA model is about 52.81%, PC1 accounts for 39.02% of the total variable, and PC2 accounts for 13.79% of the total variable. By further observing fig. 1, it was found that: the first-class wine (A), the top-class wine (B) and the special-class wine (C) are distributed relatively intensively, but the PCA analysis model cannot distinguish Fenjiu base wine samples with different sensory grades, the outlier distribution of individual top-class wine and special-class wine samples is large, and the reason of the outlier fluctuation distribution is to be further traced.
2.2.2 Linear Discriminant Analysis (LDA)
In order to solve the problem that the PCA method cannot distinguish samples of the Fenjiu base wine with different sensory grades, an LDA method is introduced for further data analysis. Due to the fact that1The raw data of the H NMR spectrum has differential variables which are effective to classification and simultaneously contains a plurality of variables which are not effective to classificationThere is also a problem of high co-linearity. To solve this problem, the present invention employs the interpretation of PCA1The 8 principal components independent of each other, determined by a 95.9% variance within the H NMR data set, were used as input variables for LDA. The linear discriminant function established by PCA/LDA can maximize the effect of the classification and visualization of the data with high dimension (multivariable) and small samples. The linear discriminant functions 1 and 2 of the PCA/LDA model of the sensory rating of Fenjiu base wine established by the invention have scoring effects, as shown in figure 2. It can be seen from FIG. 2 that the degree of dispersion in the first class (A), the second class (B) and the third class (C) samples is reduced, and the distinction between the groups is clearer.
2.3 model validation and evaluation
In the actual data analysis process, the model is easy to generate an overfitting phenomenon, so that a good false image of classification discrimination effect is obtained, and therefore, the verification of the multivariate statistical model is very important for evaluating the effectiveness of the model. Therefore, the invention respectively adopts an internal one-left cross validation method and an external repeated double-random cross validation method to validate the stability, the accuracy and the reliability of the classification result of the PCA/LDA model. FIG. 3 clearly and intuitively explains the key steps of two verification methods [ Fan S, Zhong Q, Fauhl-Hassek C, et al1H NMR spectroscopy combined with multivariate statistical analysis[J].Food Control,2018.]。
In the first step, a method of internal one-out-of-one cross validation is adopted to evaluate the performance of the PCA/LDA model. Principle of leave-one-out cross-validation method: assuming that the data set has N samples, randomly selecting N-1 samples as training samples, using the remaining sample as a test sample, and circularly obtaining N classification results for N times in sequence to be used for conservatively measuring the classification performance of the model. Table 1 summarizes the results of classifying the entire data sets of fen wine base first class (a), top class (B) and top class (C) wines respectively using the PCA/LDA model with internal cross validation. As can be seen from Table 1, after the internal leave-one cross validation, the correct classification rates of the PCA/LDA model for the first-class wine (A), the top-class wine (B) and the special-class wine (C) are respectively 95%, 88.8% and 92.8%, and the average correct recognition rate is 92.2%. The results show that1H NMR techniqueThe characteristic information of the organic matters of the Fenjiu base liquor is detected by the operation, and the PCA/LDA model is combined, so that effective distinguishing of Fenjiu base liquor samples with different sensory grades can be well realized, and the remarkable difference of the Fenjiu base liquor samples with different sensory grades on trace organic components is further shown.
Table 1 internal leave-one-out cross-validated PCA/LDA model Classification results for Fenjiu base grade sample whole data set
Second step ofMethod for cross validation using repeated double randomThe performance of the PCA/LDA model was evaluated. The main principle of the repeated double-random cross validation is that firstly, a random layered sampling mode (note: the invention carries out random layered repeated sampling for 100 times) is adopted, and an original data set is divided into two parts, namely (1) 70% of data is used as a training set for modeling; (2) 30% of the data was used as the external validation set for prediction. Since the entire data set was randomly sampled 100 times for the outer validation set, 100 different outer validation sets were obtained. The mean, standard deviation, median and absolute standard deviation of the median of the correct classification rates of the 100 samples of the external validation set were statistically analyzed (see table 2), and the 4 indices were used to evaluate the stability, validity and reliability of the model. The more the two index values of the average value and the median of the correct classification rate are close to 1, the smaller the standard deviation and the absolute standard deviation value of the median are, and the stronger the stability of the correct classification capability of the established model is.
As can be seen from Table 2, after the PCA/LDA is subjected to external repeated double random cross validation, the mean value and the median of the correct classification rates of the Fenjiu first-class wine (A), the superior-class wine (B) and the special-class wine (C) are all more than or equal to 0.74, and the mean value and the median have no significant difference (P)>0.05), and the standard deviation corresponding to the mean value and the median is less than +/-10%, and the data fluctuation is small. After repeated double random cross validation, based on1The PCA/LDA model of the H NMR technology can well realize the average correct classification of Fenjiu first-class wine (A), superior-class wine (B) and superior-class wine (C) samples of more than 82 percent, and the result shows that the PCThe effectiveness of the A/LDA model for grading evaluation on Fenjiu-based sensory quality grades.
TABLE 2 mean, standard deviation, median and median absolute deviation of correct classification rates for the repeated double random cross-validation PCA/LDA model
In summary, the present invention is achieved by non-targeting1The H NMR technology is combined with a multivariate statistical method, and the sensory intelligentized and digital graded evaluation of the fen-liquor base liquor with the faint scent is realized for the first time at home and abroad. The results show that1The PCA/LDA model of the H NMR technology has the average accuracy of 92.2 percent for grading the Fenjiu base wine, and the digital grading evaluation effect is good; the effectiveness of the PCA/LDA model is fully discussed, verified and evaluated by a one-in-one cross validation method and a repeated double-random cross validation method, and a good working foundation is laid for constructing a fen-liquor sensory intelligentized and digitized graded evaluation software system. The invention is beneficial to improving the quality of Fenjiu products, deepens the research and application of modern instrument analysis technology in the field of intelligent sensory evaluation of white spirit, promotes the progress of intelligent manufacture of Chinese white spirit industry and assists high-quality healthy sustainable development of the white spirit industry.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. An intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor comprises the following steps:
(1) using one-dimensional nuclear magnetic resonance hydrogen spectrum (1H NMR) technology to determine the characteristic fingerprint information of hydrogen atoms of organic matters in the fen-liquor base liquor to obtain a spectrogram;
(2) performing spectrum pretreatment and segmentation integration on the spectrogram;
(3) carrying out data preprocessing on the data obtained by the segmented integration;
(4) establishing a fen-flavor Fenjiu sensory quality grade and a fen-flavor Fenjiu sensory grade PCA/LDA model of data obtained by the segmented integration in the step (3) by adopting a multivariate statistical method combining unsupervised Principal Component Analysis (PCA) and supervised Linear Discriminant Analysis (LDA);
(5) verifying the PCA/LDA model obtained in the step (4) by adopting an internal one-left cross verification method and an external repeated double-random cross verification method;
(6) evaluation of the PCA/LDA model described above: and evaluating the correct classification capability and stability of the PCA/LDA model through the average value, the standard deviation, the median and the absolute standard deviation of the median of the correct classification rates of the prediction grade and the actual sensory quality grade of the fen-liquor base liquor sensory grade PCA/LDA model.
2. The method of claim 1,
in the step (2), the atlas preprocessing comprises Fourier transformation, baseline correction, phase adjustment, chemical shift calibration and regional segmentation integral correction.
3. The method of claim 1,
in the step (3), the data preprocessing includes zero-mean normalization of the data.
4. The method of claim 1,
in the step (4), when the sensory grade PCA/LDA model of the fen-flavor Fenjiu base liquor is established, the fen-flavor Fenjiu base liquor is used1Taking segmented integral Bins data of the H NMR spectrum as an independent variable X matrix, and establishing an unsupervised PCA model;
selecting a principal component score matrix with the variable accumulation variance contribution rate of more than 95% in the PCA model as an input variable, selecting the sensory quality grade of the fen-liquor base liquor with the fen-flavor type established by a sensory evaluation team as a classification standard of the supervised LDA model, and establishing the supervised LDA model.
5. The method of claim 1,
in the step (5), the repeated double-random cross validation specifically includes selecting a random hierarchical sampling mode, randomly extracting 70% of data in an original data set as a training set for modeling, 30% of data as an external validation set for prediction, and repeatedly sampling 100 times to validate a supervised LDA model.
6. The method of claim 1,
in the step (6), the more the average value of the correct classification rate is close to 1, the more the median is close to 1, the smaller the standard deviation value is, the smaller the absolute standard deviation value of the median is, and the stronger the stability of the correct classification capability of the model is.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111038566.XA CN113935573A (en) | 2021-09-06 | 2021-09-06 | Intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with delicate fragrance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111038566.XA CN113935573A (en) | 2021-09-06 | 2021-09-06 | Intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with delicate fragrance |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113935573A true CN113935573A (en) | 2022-01-14 |
Family
ID=79275077
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111038566.XA Pending CN113935573A (en) | 2021-09-06 | 2021-09-06 | Intelligent and digital grading evaluation method for sensory quality of fen-liquor base liquor with delicate fragrance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113935573A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107238638A (en) * | 2017-06-28 | 2017-10-10 | 四川理工学院 | The assay method contacted based on each composition physical and chemical index of Daqu and liquor output and vinosity |
| CN113138181A (en) * | 2021-04-27 | 2021-07-20 | 江南大学 | Method for grading quality of fresh scent type raw wine |
-
2021
- 2021-09-06 CN CN202111038566.XA patent/CN113935573A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107238638A (en) * | 2017-06-28 | 2017-10-10 | 四川理工学院 | The assay method contacted based on each composition physical and chemical index of Daqu and liquor output and vinosity |
| CN113138181A (en) * | 2021-04-27 | 2021-07-20 | 江南大学 | Method for grading quality of fresh scent type raw wine |
Non-Patent Citations (5)
| Title |
|---|
| SHUANGXI FAN, ET AL.: "Classification of Chinese wine varieties using 1H NMR spectroscopy combined with multivariate statistical analysis", 《FOOD CONTROL》, vol. 88, 4 November 2017 (2017-11-04), pages 114 - 121 * |
| 刘威 等: "核磁共振波谱技术在食品掺假鉴别中的应用研究", 《食品安全质量检测学报》, vol. 7, no. 11, 25 November 2016 (2016-11-25), pages 4358 - 4363 * |
| 张鑫: "电子鼻分析技术在白酒品质检定中的应用研究", 《酿酒》, vol. 41, no. 2, 20 March 2014 (2014-03-20), pages 27 - 29 * |
| 栾晓菲 等: "基于核磁共振技术结合多元统计分析的啤酒化学组分定量与品牌鉴别", 《酿酒科技》, no. 05, 18 May 2016 (2016-05-18), pages 48 - 53 * |
| 樊双喜 等: "基于非目标1H NMR指纹图谱技术验证中国葡萄酒原产地", 《食品与发酵工业》, vol. 44, no. 2, 25 February 2018 (2018-02-25), pages 188 - 191 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cynkar et al. | Classification of Tempranillo wines according to geographic origin: Combination of mass spectrometry based electronic nose and chemometrics | |
| He et al. | Rapid analysis of Baijiu volatile compounds fingerprint for their aroma and regional origin authenticity assessment | |
| del Mar Contreras et al. | A robustness study of calibration models for olive oil classification: Targeted and non-targeted fingerprint approaches based on GC-IMS | |
| Cheng et al. | Determination of Chinese liquors from different geographic origins by combination of mass spectrometry and chemometric technique | |
| Mascellani et al. | 1H NMR chemometric models for classification of Czech wine type and variety | |
| Solovyev et al. | NMR spectroscopy in wine authentication: An official control perspective | |
| Roullier-Gall et al. | How subtle is the “terroir” effect? Chemistry-related signatures of two “climats de Bourgogne” | |
| Gishen et al. | Grape and wine analysis‐enhancing the power of spectroscopy with chemometrics. A review of some applications in the Australian wine industry | |
| Li et al. | Vintage analysis of Chinese Baijiu by GC and 1H NMR combined with multivariable analysis | |
| Bajoub et al. | A metabolic fingerprinting approach based on selected ion flow tube mass spectrometry (SIFT-MS) and chemometrics: A reliable tool for Mediterranean origin-labeled olive oils authentication | |
| Martínez-Jarquín et al. | Automated chemical fingerprinting of Mexican spirits derived from Agave (tequila and mezcal) using direct-injection electrospray ionisation (DIESI) and low-temperature plasma (LTP) mass spectrometry | |
| Cynkar et al. | Feasibility study on the use of a head space mass spectrometry electronic nose (MS e_nose) to monitor red wine spoilage induced by Brettanomyces yeast | |
| Papotti et al. | Traditional balsamic vinegar and balsamic vinegar of Modena analyzed by nuclear magnetic resonance spectroscopy coupled with multivariate data analysis | |
| Belmonte-Sánchez et al. | An innovative metabolomic approach for golden rum classification combining ultrahigh-performance liquid chromatography–orbitrap mass spectrometry and chemometric strategies | |
| He et al. | Comprehensive two-dimensional gas chromatography–time of flight mass spectrometry (GC× GC-TOFMS) in conventional and reversed column configuration for the investigation of Baijiu aroma types and regional origin | |
| CN111521722A (en) | Method for identifying storage years of fragrant odor type finished product white spirit bottles | |
| Tzachristas et al. | LC-MS based metabolomics for the authentication of selected Greek white wines | |
| De Carvalho et al. | Determination of the geographical origin and ethanol content of Brazilian sugarcane spirit using near-infrared spectroscopy coupled with discriminant analysis | |
| Ceballos-Magana et al. | Quantitation of twelve metals in tequila and mezcal spirits as authenticity parameters | |
| Stockwell et al. | Quantification of whisky congeners by 1H NMR spectroscopy | |
| CN109214423B (en) | Food quality discrimination analysis method based on dynamic and static data fusion | |
| Okolo et al. | Recent advances in whiskey analysis for authentication, discrimination, and quality control | |
| Hayashi et al. | 1H-NMR metabolomics-based classification of Japanese sake and comparative metabolome analysis by gas chromatography–mass spectrometry | |
| Aurum et al. | Coffee origin determination based on analytical and nondestructive approaches–A systematic literature review | |
| Wu et al. | Geographical origin traceability and authenticity detection of Chinese red wines based on excitation-emission matrix fluorescence spectroscopy and chemometric methods |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |