Cigar tobacco leaf sensory quality prediction method

Abstract

The invention belongs to the technical field of cigar tobacco analysis, and particularly relates to a cigar tobacco sensory quality prediction method. The method comprises the following steps: S1, obtaining content data of various chemical components of cigar tobacco leaves to be detected; the chemical components comprise volatile chemical components and non-volatile chemical components; and S2, outputting the content data into a pre-trained sensory quality prediction model to obtain a sensory score of the cigar tobacco to be detected. Compared with the prior art in which the sensory quality of the tobacco leaves is predicted only according to the non-volatile chemical components of the to-be-detected tobacco leaves, the method has the advantages that the content of various volatile chemical components is comprehensively considered to predict the sensory quality of the tobacco leaves, so that the characterization effect of the volatile chemical components on the sensory quality of the sample is greatly exerted; and the accuracy and rationality of sensory quality evaluation of cigars are greatly improved. The technical problem that in the prior art, sensory prediction is carried out only according to non-volatile chemical components, and it is difficult to accurately evaluate the sensory quality of cigars is solved.

Description

Technical Field

[0001] This invention belongs to the field of sensory testing technology for cigar tobacco leaves, and specifically relates to a method for predicting the sensory quality of cigar tobacco leaves. Background Technology

[0002] Cigars, as an important tobacco product, are renowned for their rich aroma, alkaline smoke, complex flavor, and low tar-to-nicotine ratio. Different genotypes, growing environments, agricultural practices, and processing techniques, among other factors, endow cigars with unique chemical compositions and physical properties, thus influencing their quality. Cigar quality is often assessed using traditional sensory evaluation methods, relying on smell, taste, and the quality of the smoke produced when the cigar tobacco is burned. However, traditional sensory evaluation methods are highly subjective, easily influenced by personal preferences and mood, and are also time-consuming and labor-intensive.

[0003] A Chinese invention patent application with publication number CN116660458A and publication date of August 29, 2023, discloses a method for predicting the sensory quality of cigar raw materials based on a backpropagation (BP) neural network. This method obtains the sensory quality of the cigar tobacco leaves by training a BP neural network based on conventional chemical components such as total nitrogen and total sugar. However, this method only predicts sensory quality based on conventional non-volatile chemical components. While the measurement of non-volatile chemical components is a commonly used parameter for determining the quality of cigar tobacco leaves and can provide important information for quality measurement and control, volatile components also have a significant characterizing effect on the sensory quality of the sample. Therefore, relying solely on non-volatile chemical components for sensory prediction makes it difficult to accurately assess the sensory quality of cigars. Summary of the Invention

[0004] The purpose of this invention is to provide a method for predicting the sensory quality of cigar tobacco leaves, in order to solve the technical problem that the existing technology, which relies solely on non-volatile chemical components for sensory prediction, makes it difficult to accurately assess the sensory quality of cigars.

[0005] To solve the above-mentioned technical problems, the present invention provides a method for predicting the sensory quality of cigar tobacco leaves, comprising: S1. Obtain the content data of various chemical components in the cigar tobacco leaves to be tested; the chemical components include volatile chemical components and non-volatile chemical components; S2. Input the content data into a pre-trained sensory quality prediction model to obtain the sensory score of the cigar tobacco leaf to be tested.

[0006] The beneficial effects of the above technical solution are as follows: The technical solution of the cigar tobacco leaf sensory quality prediction method of the present invention belongs to an improved invention. Unlike the prior art that only predicts the sensory quality of tobacco leaves based on the non-volatile chemical components, the present invention comprehensively considers the content of multiple volatile chemical components to predict the sensory quality of tobacco leaves, greatly leveraging the characterizing role of volatile chemical components on the sensory quality of samples, and significantly improving the accuracy and rationality of cigar sensory quality assessment. The present invention solves the technical problem in the prior art that it is difficult to accurately assess the sensory quality of cigars when relying solely on non-volatile chemical components for sensory prediction.

[0007] Furthermore, the content data of various volatile chemical components were obtained using the GC-MS method.

[0008] Furthermore, the sensory rating includes a sensory rating of at least one of the following sensory indicators: cleanliness, fineness, aroma quantity, maturity, richness, smoothness, sweetness, irritation, aftertaste, flammability, grayness, and ashiness.

[0009] Furthermore, the sensory rating also includes a total sensory quality score derived from sensory ratings of all sensory indicators.

[0010] Furthermore, apart from the irritant sensory index, the higher the sensory score of the other sensory indexes, the more prominent the flavor characteristics corresponding to that sensory index; the higher the sensory score of the irritant sensory index, the weaker the irritation.

[0011] Furthermore, the non-volatile chemical components include conventional chemical components, organic acids, plastid pigments, amino acids, polyphenolic compounds, inorganic anions, inorganic cations, alkaloids, TSNAs, heavy metal elements, and ammonia.

[0012] Furthermore, the conventional chemical components include total sugar, reducing sugar, total nitrogen, potassium, chlorine, and protein; The organic acids include oxalic acid, tartaric acid, formic acid, malic acid, malonic acid, α-ketoglutaric acid, lactic acid, acetic acid, citric acid, maleic acid, fumaric acid, and succinic acid. The plastid pigments include neoxanthin, violetin, lutein, chlorophyll B, chlorophyll A, and β-carotene; The amino acids include aspartic acid, serine, glutamic acid, glycine, histidine, arginine, threonine, alanine, proline, 4-aminobutyric acid, cystine, tyrosine, valine, methionine, lysine, isoleucine, leucine, and phenylalanine. The polyphenolic compounds include neochlorogenic acid, chlorogenic acid, cryptochlorogenic acid, hyoscyamine, and rutin; The inorganic anions include nitrate, phosphate, and sulfate. The inorganic cations include Na ions, Fe ions, Mg ions, K ions, and Ca ions; The alkaloids include nicotine, nornicotine, mesmin, pseudoestiline, β-dienenicotine, neonicotine, 2,3-bipyridine, and cotinine; The TSNAs include NNN, NNK, NAT, and NAB; The heavy metal elements include Cr, Ni, Cu, Zn, As, Se, Cd, Pb, Sn, and Sb.

[0013] Furthermore, the sensory quality prediction model is the XGBoost model. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating an implementation method for predicting the sensory quality of cigar tobacco leaves according to the present invention. Detailed Implementation

[0015] Unlike existing technologies that predict the sensory quality of tobacco leaves solely based on their non-volatile chemical components, this invention comprehensively considers the content of multiple volatile chemical components to predict the sensory quality of tobacco leaves. This greatly leverages the characterizing role of volatile chemical components in the sensory quality of samples, significantly improving the accuracy and rationality of cigar sensory quality assessment. This invention solves the technical problem of existing technologies that rely solely on non-volatile chemical components for sensory prediction, making it difficult to accurately assess the sensory quality of cigars.

[0016] Implementation method for predicting the sensory quality of cigar tobacco leaves: A method for predicting the sensory quality of cigar tobacco leaves, such as Figure 1 As shown, it includes the following steps: Step 1: Obtain data on the volatile components of cigar tobacco leaves.

[0017] This implementation method collects volatile component data from 135 cigar tobacco samples to provide a basis for constructing a sensory evaluation prediction model.

[0018] The data collection steps for volatile components of cigar tobacco leaves were as follows: 27 samples from Yunnan, 30 from Sichuan, 32 from Hainan, and 46 from Hubei were collected. 2.0 g of cigar tobacco powder was accurately weighed and placed in a 50 mL centrifuge tube. 10 mL of water was added and the mixture was soaked for 10 min. Then, 10 mL of acetonitrile and 0.1 mL of internal standard solution (phenethyl acetate, concentration 100 µg / mL) were added. The mixture was vortexed for 1 min and then frozen for 10 min. A high-performance sample extraction kit was then added, and the mixture was vortexed for 2 min to prevent magnesium sulfate from clumping. After centrifugation at 4000 r / min for 10 min, 1 mL of the supernatant was transferred to a 1.5 mL centrifuge tube. A high-performance pigment sample purification kit was added, and the mixture was vortexed at 2000 r / min for 2 min, then centrifuged at 6000 r / min for 2 min. The supernatant was collected for GC-MS analysis.

[0019] The GC-MS analysis conditions were as follows: a DB-5 quartz capillary column (60 m × 0.25 mm × 0.25 µm) was used, with an injection port temperature of 270 ℃, a split ratio of 20:1, and an injection volume of 1 µL. Helium was used as the carrier gas, and the flow rate was constant at 1 mL / min with a solvent delay time of 6 min. The temperature program was as follows: initial temperature of 60 ℃ held for 0.5 min, then increased to 280 ℃ at a rate of 2 ℃ / min and held for 30 min. The ion source was an electron impact (EI) source with an ionization voltage of 70 eV, an ion source temperature of 230 ℃, and a transfer line temperature of 280 ℃. The mass scan range was 30–350 amu, and data were acquired simultaneously using full scan and selected ion monitoring (SIM) modes.

[0020] Step 2: Obtain data on the non-volatile chemical composition of cigar tobacco leaves.

[0021] This embodiment collects non-volatile chemical composition data of 135 cigar tobacco leaves mentioned in step 1 to provide a basis for the subsequent construction of a sensory evaluation prediction model.

[0022] The data on non-volatile chemical components include the content of conventional chemical components, organic acids, plastid pigments, amino acids, polyphenolic compounds, inorganic anions, inorganic cations, alkaloids, TSNAs, heavy metals, and ammonia in each sample.

[0023] The conventional chemical components include total sugar, reducing sugar, total nitrogen, potassium, chlorine, and protein.

[0024] Organic acids include oxalic acid, tartaric acid, formic acid, malic acid, malonic acid, α-ketoglutaric acid, lactic acid, acetic acid, citric acid, maleic acid, fumaric acid, and succinic acid.

[0025] Plastid pigments include neoxanthin, violetin, lutein, chlorophyll B, chlorophyll A, and β-carotene.

[0026] Amino acids include aspartic acid, serine, glutamic acid, glycine, histidine, arginine, threonine, alanine, proline, 4-aminobutyric acid, cystine, tyrosine, valine, methionine, lysine, isoleucine, leucine, and phenylalanine.

[0027] Polyphenolic compounds include neochlorogenic acid, chlorogenic acid, cryptochlorogenic acid, hyoscyamine, and rutin.

[0028] Inorganic anions include nitrate, phosphate, and sulfate.

[0029] Inorganic cations include Na ions, Fe ions, Mg ions, K ions, and Ca ions.

[0030] Alkaloids include nicotine, nornicotine, mesmin, pseudoestipine, β-dienenicotine, neonicotine, 2,3-bipyridine, and cotinine.

[0031] TSNAs (Tobacco-Specific Nitrosamines) include NNN, NNK, NAT, and NAB.

[0032] Heavy metal elements include Cr, Ni, Cu, Zn, As, Se, Cd, Pb, Sn, and Sb.

[0033] The data collection methods for non-volatile chemical components of cigar tobacco leaves are as follows: Gas chromatography-mass spectrometry (GC-MS) was used to determine acidic substances in cigar tobacco leaves; plastid pigments were determined according to industry standard YC / T 382-2010; amino acids and polyphenols were determined according to high-performance liquid chromatography (HPLC); ammonia was determined according to enterprise standard YQ / T19-2012; inorganic anions were determined according to ion chromatography; inorganic cations were determined according to inductively coupled plasma mass spectrometry (ICP-MS); alkaloids were determined according to industry standard YC / T 383-2010; total nitrogen, total sugar, reducing sugar, chlorine, and potassium were determined according to industry standards using continuous flow chromatography; and heavy metals were determined using ICP-MS.

[0034] Step 3: Obtain sensory rating data for cigar tobacco leaves.

[0035] This embodiment collects sensory evaluation data from 135 cigar tobacco samples described in step 1 to provide a basis for constructing a sensory evaluation prediction model.

[0036] The sensory evaluation dataset for cigar tobacco leaves was collected following steps: Cigar tobacco leaf samples were hand-rolled by professional rollers into cigars 110 mm in length and 14 mm in diameter, and then equilibrated for 72 hours at 22°C and 65% relative humidity to stabilize their moisture content. Sensory quality assessment was conducted by a panel of five experienced experts, all with over five years of sensory evaluation experience and more than six months of professional training. The assessment was based on 12 sensory indicators: cleanliness, fineness, aroma intensity, maturity, richness, smoothness, sweetness, irritation, finish, combustibility, ash, and ash density. Scoring ranged from 0 to 9 points; except for irritation indicators, higher scores for the other sensory indicators indicated more prominent flavor characteristics, while higher scores for irritation indicators indicated less irritation. The score for each sample was unanimously approved by all expert panel members. In addition to the 12 sensory indicators, a total sensory quality score of 100 points was also included, representing the overall score of the sample based on the 12 sensory indicators.

[0037] Step 4: The dataset is processed using the machine learning model XGBoost, with the content data of volatile components and non-volatile chemical components as input variables and sensory scores as output variables to build a predictive model.

[0038] Step 5: Use Leave-One-Out Cross-Validation (LOOCV) to evaluate model performance.

[0039] Leave-one-out cross-validation was used to evaluate model performance, and the root mean square error (RMSE) was compared with the standard deviation (SD) to analyze prediction accuracy.

[0040] The model with the highest prediction accuracy is selected as the final sensory quality prediction model. The content data of various chemical components of the cigar tobacco leaves to be tested are input into the obtained sensory quality prediction model. The sensory quality prediction model can then output the sensory scores of each sensory indicator and the total sensory quality score. Except for the irritant sensory indicators, the higher the sensory score of the other sensory indicators, the better and more prominent the flavor characteristics corresponding to that sensory indicator of the cigar tobacco leaves to be tested are; the higher the score of the irritant sensory indicator, the weaker the irritation.

[0041] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still make modifications to the technical solutions described in the foregoing embodiments without creative effort, or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the sensory quality of cigar tobacco leaves, characterized in that, The method includes: S1. Obtain the content data of various chemical components in the cigar tobacco leaves to be tested; the chemical components include volatile chemical components and non-volatile chemical components; S2. Input the content data into a pre-trained sensory quality prediction model to obtain the sensory score of the cigar tobacco leaf to be tested.

2. The method for predicting the sensory quality of cigar tobacco leaves according to claim 1, characterized in that, The content data of various volatile chemical components were obtained using the GC-MS method.

3. The method for predicting the sensory quality of cigar tobacco leaves according to claim 1, characterized in that, The sensory rating includes a sensory rating of at least one of the following sensory indicators: cleanliness, fineness, aroma quantity, maturity, richness, smoothness, sweetness, irritation, aftertaste, flammability, grayness, and ashiness.

4. The method for predicting the sensory quality of cigar tobacco leaves according to claim 3, characterized in that, The sensory rating also includes a total sensory quality score derived from the sensory ratings of all sensory indicators.

5. The method for predicting the sensory quality of cigar tobacco leaves according to claim 3 or 4, characterized in that, Except for irritant sensory indicators, the higher the sensory score of other sensory indicators, the more prominent the flavor characteristics corresponding to that sensory indicator; the higher the sensory score of irritant sensory indicators, the weaker the irritation.

6. The method for predicting the sensory quality of cigar tobacco leaves according to claim 1, characterized in that, The non-volatile chemical components include conventional chemical components, organic acids, plastid pigments, amino acids, polyphenolic compounds, inorganic anions, inorganic cations, alkaloids, TSNAs, heavy metal elements, and ammonia.

7. The method for predicting the sensory quality of cigar tobacco leaves according to claim 6, characterized in that, The conventional chemical components include total sugar, reducing sugar, total nitrogen, potassium, chlorine, and protein; The organic acids include oxalic acid, tartaric acid, formic acid, malic acid, malonic acid, α-ketoglutaric acid, lactic acid, acetic acid, citric acid, maleic acid, fumaric acid, and succinic acid. The plastid pigments include neoxanthin, violetin, lutein, chlorophyll B, chlorophyll A, and β-carotene; The amino acids include aspartic acid, serine, glutamic acid, glycine, histidine, arginine, threonine, alanine, proline, 4-aminobutyric acid, cystine, tyrosine, valine, methionine, lysine, isoleucine, leucine, and phenylalanine. The polyphenolic compounds include neochlorogenic acid, chlorogenic acid, cryptochlorogenic acid, hyoscyamine, and rutin; The inorganic anions include nitrate, phosphate, and sulfate. The inorganic cations include Na ions, Fe ions, Mg ions, K ions, and Ca ions; The alkaloids include nicotine, nornicotine, mesmin, pseudoestiline, β-dienenicotine, neonicotine, 2,3-bipyridine, and cotinine; The TSNAs include NNN, NNK, NAT, and NAB; The heavy metal elements include Cr, Ni, Cu, Zn, As, Se, Cd, Pb, Sn, and Sb.

8. The method for predicting the sensory quality of cigar tobacco leaves according to claim 1, characterized in that, The sensory quality prediction model is the XGBoost model.

Images