Method for establishing moistening base module based on machine learning

Abstract

The invention relates to a method for establishing a moistening base module based on machine learning, and the method comprises the steps: selecting a moistening function formula, constructing a perfume formula moistening function sensory evaluation method, and obtaining and verifying data of a moistening core perfume raw material and an auxiliary perfume raw material; selecting a representative moistening function formula, and by taking the representative moistening function formula as a reference, imitating a moistening core incense raw material and an imitating function formula by utilizing data of the moistening core incense raw material and the auxiliary incense raw material; and inputting the representative moistening function formula and the component content and moistening feeling data of the imitation function formula into a pre-trained machine learning model to obtain a created formula. The invention provides a machine learning-based moistening base module creation method, which lays a foundation for functional analysis and digital research of tobacco flavor formula components and provides a new thought for tobacco flavor raw material perfuming control and cigarette product quality improvement.

Description

Technical Field

[0001] The present application belongs to the technical field of cigarette product design, and specifically, relates to a method for establishing a moisturizing aroma base module based on machine learning. Background Art

[0002] As an important raw material in the production of cigarettes, the formulation research and development of tobacco flavors involves the core technology of the tobacco industry. At present, most of the research on the creation of flavor base modules still relies mainly on sensory evaluation and empirical rules, which is labor-intensive, time-consuming and labor-intensive. However, the raw materials of tobacco flavor base modules are of many types and complex ingredients. The relationship between many raw materials and sensory functions is unclear, making it difficult to achieve targeted flavor enhancement and characteristic creation.

[0003] Therefore, it is an urgent problem for technical personnel in this field to clarify the relationship between tobacco flavor base module raw materials and sensory functions, develop new methods for creating flavor base modules, and provide new research ideas for the digital research of tobacco flavor raw materials and the improvement of cigarette product quality. Summary of the invention

[0004] This application is based on at least one of the above-mentioned technical problems, and proposes a method for creating a moisturizing flavor base module based on machine learning. The idea is to clarify the moisturizing core flavor raw materials and auxiliary flavor raw materials in the formula through sensory evaluation, formula data analysis, imitation verification and machine learning model construction, and to use machine learning to construct a mathematical model through moisturizing score values ​​and the content of monomer flavors in the formula. The machine learning model outputs the created formula and performs sensory verification. This invention lays the foundation for the functional analysis and digital research of tobacco flavor formula ingredients, and provides new ideas for the control of flavoring tobacco flavor raw materials and the improvement of cigarette product quality.

[0005] To achieve the above objectives, the present application provides a method for establishing a moisturizing fragrance base module based on machine learning, selecting a moisturizing functional formula and constructing a sensory evaluation method for the moisturizing function of a fragrance formula, obtaining and verifying the data of moisturizing core fragrance raw materials and auxiliary fragrance raw materials;

[0006] Select representative moisturizing functional formulas and use them as a benchmark, and use the data of moisturizing core fragrance raw materials and auxiliary fragrance raw materials to imitate formulas containing moisturizing core fragrance raw materials and imitation functional formulas;

[0007] The ingredient content and moisturizing data of the representative moisturizing functional formula and the imitation functional formula are input into a pre-trained machine learning model to obtain a creative formula.

[0008] In some embodiments, the method for establishing a pre-trained machine learning model comprises the following steps:

[0009] With the lubrication function fitting output value Y and different monomer component contents X, the mathematical model of the lubrication function output value Y and X is calculated by machine learning fitting;

[0010] The machine learning algorithm can be one or more of the random forest algorithm, K nearest neighbor regression algorithm, support vector machine algorithm, decision tree algorithm, etc., and the R 2 Greater than 0.80.

[0011] In some embodiments, a method for constructing a sensory evaluation method for the moisturizing function of a fragrance formula is used to select a fragrance formula with a moisturizing function, and a sensory evaluation of the moisturizing function is performed and the scores are summarized.

[0012] In some embodiments, constructing a sensory evaluation method for the moisturizing function of a fragrance formula comprises:

[0013] The nasal moistness evaluation score and the tongue surface salivation evaluation score are set separately; and the sum of the nasal moistness evaluation score and the tongue surface salivation evaluation score is taken as the total score; the moistness score is the difference between the moistness scores of the test cigarette and the blank cigarette;

[0014] The total score and moistness score were used to construct the moistness sensory evaluation data.

[0015] In some embodiments, the similarities in the composition of formula raw materials whose moistness scores reach a set value are grouped and the core fragrance raw material monomers and auxiliary fragrance raw material monomers are clearly defined, and the core fragrance raw material monomers and auxiliary fragrance raw material monomers are evaluated and verified separately using a sensory evaluation method for constructing the moisturizing function of a fragrance formula.

[0016] In some embodiments, 50-200 representative moisturizing functional formulas are selected.

[0017] In some embodiments, there are 20-50 representative moisturizing functional formulas, 1-15 moisturizing core fragrance raw material monomers and 0-10 auxiliary fragrance raw material monomers.

[0018] In some embodiments, the monomer contents of 1-5 representative moisturizing functional formulas are selected as the basis, 30-90 imitation formulas containing core fragrance raw materials and auxiliary fragrance raw materials are configured, and the moisturizing sensory evaluation data are obtained by constructing a sensory evaluation method for the moisturizing function of a fragrance formula.

[0019] In some embodiments, the number of created formulas is 1-5, and a sensory evaluation of the moisturizing function of the spice formula is performed, and the machine learning model is optimized using the sensory evaluation data of the moisturizing function of the spice formula.

[0020] This application selects representative moisturizing functional formulas, constructs a sensory evaluation method for the moisturizing function of fragrance formulas, selects moisturizing core fragrance raw materials and auxiliary fragrance raw materials through data analysis technology and other means, and verifies them. Based on the formula with better moisturizing effect, the formula containing core fragrance raw materials and auxiliary fragrance raw materials is imitated. The representative moisturizing functional formula and the imitated functional formula are based on the ingredient content and moisturizing data, and a model is established through training of machine learning algorithms such as neural networks to clarify the relationship between moisturizing and the content of formula ingredients. The imitated formula is output according to the machine learning model, and verification and model optimization are performed. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of a random forest model fitting provided in one embodiment of the present application;

[0022] Figure 2 A schematic diagram of support vector machine model fitting provided in one embodiment of the present application;

[0023] Figure 3 A schematic diagram of model fitting of a K nearest neighbor regression model provided in an embodiment of the present application;

[0024] Figure 4 A schematic diagram of model fitting of a decision tree model provided in one embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present application is not limited to the specific embodiments disclosed below.

[0027] The present application discloses a method for creating a moisturizing fragrance base module based on machine learning, selects representative moisturizing functional formulas, constructs a sensory evaluation method for the moisturizing function of fragrance formulas, screens out moisturizing core fragrance raw materials and auxiliary fragrance raw materials through data analysis technology, and verifies them, and imitates formulas containing core fragrance raw materials and auxiliary fragrance raw materials based on formulas with better moisturizing effects. The representative moisturizing functional formulas and imitated functional formulas are based on the ingredient content and moisturizing data, and a model is established through training of machine learning algorithms such as neural networks to clarify the relationship between moisturizing and the content of formula ingredients. The imitated formula is output according to the machine learning model, and verification and model optimization are performed. The specific method includes the following steps.

[0028] Step 1: Construct a sensory evaluation method for the moisturizing function of spice formulas, select spice formulas with moisturizing function, and conduct sensory evaluation of the moisturizing function and summarize the scores.

[0029] Among them, 50-200 kinds of moisturizing functional flavor formulas were selected, and the constructed sensory evaluation method for the moisturizing function of the flavor formula included nasal moisturizing (50 points) and tongue salivation (50 points), and the total score was the sum of nasal moisturizing and tongue salivation; the moisturizing functional formula was added to blank cigarettes to form test cigarettes, and 8 or more professionally trained evaluators conducted sensory evaluation on the test cigarettes according to the above sensory evaluation method, and the final moisturizing score was the difference between the moisturizing scores of the test cigarettes and the blank cigarettes.

[0030] Step 2: Use data analysis technology to group formulas with higher moisturizing function scores according to the similarity of formula raw material composition, and select 20-50 formulas with higher moisturizing function scores. Use data mining technology to analyze raw material formula data, identify the core fragrance raw material monomers and auxiliary fragrance raw material monomers that appear frequently in each group of formulas, and conduct sensory evaluation and verification of moisturizing function respectively. Analyze 1-15 core fragrance raw material monomers and 0-10 auxiliary fragrance raw material monomers that appear frequently in each group of formulas.

[0031] Step three: Based on the monomer content of formulas with better moisturizing effects, specifically based on the monomer content of 1-5 selected formulas with better moisturizing effects, prepare 30-90 imitation formulas containing core fragrance raw materials and auxiliary fragrance raw materials, and conduct moisturizing function evaluation to obtain moisturizing sensory evaluation data.

[0032] Step 4: Select the imitation formula in step 3 and the formula with higher moistening function score in step 2, take the moistening function fitting output value Y and different monomer component contents as X, and calculate the moistening function output value Y and X mathematical model through machine learning fitting.

[0033] That is, 30-90 imitation formulas of step three and 20-50 formulas with higher moisturizing function scores of step two are selected, with the final moisturizing score as the output value Y, and different monomer component contents as X1, X2, X3..., and the mathematical model of the moisturizing function output value Y and X1, X2, X3... is calculated through machine learning fitting.

[0034] The machine learning algorithm can be one or more of the random forest algorithm, K nearest neighbor regression algorithm, support vector machine algorithm, decision tree algorithm, etc. 2 Greater than 0.80.

[0035] Step 5: Based on the required smoothness score, use the machine learning model to output the creation formula and conduct sensory evaluation verification. Use the smoothness score of the output creation formula to optimize and improve the model.

[0036] For example, use the machine learning model to output 1-5 creative formulas, evaluate their moisturizing function, and use their moisturizing scores to optimize the model.

[0037] Example 1 Construction of sensory evaluation method for moisturizing function and sensory evaluation of fragrance formula

[0038] According to the sensory evaluation requirements of "YC / T 497-2014 Cigarette Chinese Cigarette Style Sensory Evaluation Method" and "YC / T496-2014 Cigarette Sensory Comfort Evaluation Method", the sensory evaluation method was adjusted, and the moistness score was divided into nasal moistness (0-50 points) and tongue salivation (0-50 points), and the total moistness score was the sum of the two scores (Table 1). With reference to the moistness score gradient of common commercial finished cigarettes (Table 2), the score of the blank cigarette was determined, and 150 flavor formulas were sensory evaluated, and their moistness data were analyzed and statistically analyzed.

[0039] Table 1 Sensory evaluation method for moisturizing function

[0040]

[0041] Table 2 Summary of the smoothness scores of commercially available finished cigarettes

[0042]

[0043] Example 2 Screening of formulas with better moisturizing function

[0044] The moistening data of 150 flavor formulas were analyzed and counted, and 20 formulas with better moistening function, PF1-PF20, were selected (as shown in Table 3). The final moistening score Y is the difference between the moistening scores of the test cigarette and the blank cigarette. The raw material composition of the 20 groups of formulas was analyzed, and the ones with the highest frequency in each group of formulas were identified as core fragrance raw material monomers and auxiliary fragrance raw material monomers. Standard products were purchased and added to blank cigarettes for sensory evaluation verification. It was finally determined that the 7 core fragrance raw material monomers were lactic acid, acetic acid, phenylacetic acid, menthol, sorbitol, furanone, and ethyl butyrate, and the 4 auxiliary fragrance raw material monomers were Zimbabwe tobacco extract, fig extract, sour plum extract, and imported apricot extract.

[0045] Table 3 Summary of scores of 20 formulas with better moisturizing function

[0046] Nasal moisturizing Saliva production on tongue Score Y Blank smoke 20 20 0 PF-1 39 39 38 PF-2 40 39 39 PF-3 39 37 36 PF-4 42 43 45 PF-5 44 44 48 PF-6 41 42 43 PF-7 38 40 38 PF-8 43 42 45 PF-9 42 41 43 PF-10 42 43 45 PF-11 43 44 47 PF-12 38 40 38 PF-13 39 40 39 PF-14 36 35 31 PF-15 38 37 35 PF-16 44 43 47 PF-17 42 40 42 PF-18 43 45 48 PF-19 38 36 34 PF-20 42 41 43

[0047] Example 3: Formulation and sensory evaluation based on core fragrance raw materials and auxiliary fragrance raw materials monomers

[0048] According to the common concentration ranges of the above 7 core fragrance raw materials and 4 auxiliary fragrance raw materials in the 5 formulas, 60 imitation formulas (PF-1-PF-60) were designed, and the moistness score evaluation was carried out (Table 4) to determine the moistness score values ​​of the 60 formulas.

[0049] Table 4 Summary of moistness scores of 60 formulas

[0050]

[0051]

[0052]

[0053]

[0054]

[0055] Example 4 Machine Learning Model Construction

[0056] The 60 sets of formula data in Example 3 were collected and sorted. The smoothness score was used as the output value Y, and the contents of the 7 core fragrance raw materials and the 4 auxiliary fragrance raw materials in the formula were X (X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X12, X13, X14, X15, X16, X17, X18, X19, X20, X21, X22, X23, X24, X25, X26, X27, X28, X29, X30, X31, X32, X33, X34, X35, X36, X37, X38, X39, X40, X41, X42, X43, X44, X45, X46, X47, X48, X49, X50, X51, X52, X 10 , X 11 ), using machine learning python programming code, taking X and Y feature data as the data source, dividing them into training set and validation set in a 7:3 manner, selecting random forest model, support vector machine model, K nearest neighbor regression model, decision tree model in machine learning, using specific parameters, training set for training and fitting, validation set for validating the model, using mean absolute error (MAE) and determination coefficient R 2 As an evaluation, a machine learning model of moisturizing score and ingredient content was established. Figure 1-Figure 4 The results show that the random forest model R 2 Score 0.8921, MAE 4.3189, support vector machine model R 2 Score 0.8374, MAE 5.3487, K nearest neighbor regression model R 2 Score 0.9080, MAE 3.8333, decision tree model R 2 The score is 0.8656 and MAE is 4.8854. The best score is the K nearest neighbor regression model.

[0057] Example 5 Sensory Evaluation Verification of Model Output Formula

[0058] Assuming that the required formula has a smoothness score of 38, the random forest model is used to output two formulas with the following ratios (X1 = 0.1, X3 = 0.2, X4 = 0.002, X10 =0.02, X 11 =0.01) and (X1 = 0.2, X3 = 0.02, X6 = 0.053, X7 = 0.01, X8 = 0.01, X 10 =0.01). Two recipes were prepared according to the above ratios and injected into blank cigarettes for sensory evaluation. The results showed that the sensory evaluation mean of the first recipe was 39, and the sensory evaluation mean of the second recipe was 38, which was very close to the theoretical prediction value, indicating that the machine learning method model is relatively accurate and reliable.

[0059] In summary, this application proposes a method for creating a moisturizing flavor base module based on machine learning. The idea is to clarify the moisturizing core flavor raw materials and auxiliary flavor raw materials in the formula through sensory evaluation, formula data analysis, imitation verification, and machine learning model construction methods, and to use machine learning to construct a mathematical model through moisturizing score values ​​and the content of monomer flavors in the formula. The machine learning model outputs the created formula and performs sensory verification. This method lays the foundation for the functional analysis and digital research of tobacco flavor formula ingredients, and also provides new ideas for the control of flavoring tobacco flavor raw materials and the improvement of cigarette product quality.

[0060] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for establishing a moisturizing fragrance base module based on machine learning, characterized in that: Select moisturizing formulas and construct a sensory evaluation method for moisturizing function of fragrance formulas, obtain and verify the data of moisturizing core fragrance raw materials and auxiliary fragrance raw materials; Select representative moisturizing functional formulas and use them as a benchmark, and use the data of moisturizing core fragrance raw materials and auxiliary fragrance raw materials to imitate formulas containing moisturizing core fragrance raw materials and imitation functional formulas; The ingredient content and moisturizing data of the representative moisturizing functional formula and the imitation functional formula are input into a pre-trained machine learning model to obtain a creative formula.

2. The method according to claim 1, characterized in that The method for establishing the pre-trained machine learning model comprises the following steps: With the lubrication function fitting output value Y and different monomer component contents X, the mathematical model of the lubrication function output value Y and X is calculated by machine learning fitting; The machine learning algorithm can be one or more of the random forest algorithm, K nearest neighbor regression algorithm, support vector machine algorithm, decision tree algorithm, etc., and the R 2 Greater than 0.

80.

3. The method according to claim 1, characterized in that By constructing a sensory evaluation method for the moisturizing function of spice formulas, we selected spice formulas with moisturizing function, and conducted sensory evaluation on their moisturizing function and summarized their scores.

4. The method according to claim 3, characterized in that The sensory evaluation method for the moisturizing function of the fragrance formula includes: The nasal moistness evaluation score and the tongue surface salivation evaluation score are set separately; and the sum of the nasal moistness evaluation score and the tongue surface salivation evaluation score is taken as the total score; the moistness score is the difference between the moistness scores of the test cigarette and the blank cigarette; The total score and moistness score were used to construct the moistness sensory evaluation data.

5. The method according to claim 3, characterized in that: The formula raw material composition similarities whose moistness scores reach the set value are grouped and the core fragrance raw material monomers and auxiliary fragrance raw material monomers are clarified. The core fragrance raw material monomers and auxiliary fragrance raw material monomers are evaluated and verified separately by constructing a sensory evaluation method for the moisturizing function of the fragrance formula.

6. The method according to claim 3, characterized in that 50-200 representative formulas with moisturizing functions were selected.

7. The method according to claim 3, characterized in that There are 20-50 representative moisturizing functional formulas, 1-15 moisturizing core fragrance raw material monomers and 0-10 auxiliary fragrance raw material monomers.

8. The method according to claim 3, characterized in that Based on the monomer content of 1-5 representative moisturizing functional formulas, 30-90 imitation formulas containing core fragrance raw materials and auxiliary fragrance raw materials were configured, and the moisturizing sensory evaluation data were obtained by using the sensory evaluation method for the moisturizing function of the fragrance formula.

9. The method according to claim 3, characterized in that: The number of created formulas is 1 to 5, and a sensory evaluation of the moisturizing function of the spice formula is performed, and the machine learning model is optimized using the sensory evaluation data of the moisturizing function of the spice formula.

Images