Cigarette note characteristic similarity measurement method

Abstract

The invention provides a cigarette note characteristic similarity measurement method which comprises the following steps: selecting representative cigarettes with different brands in set brand cigarettes, organizing smoke panel test personnel to perform smoke panel test on the representative cigarettes, and recording scores of note indexes; calculating a mean value and a standard deviation of scores of the note indexes in the cigarettes of different brands, and constructing a multi-peak normal distribution model; constructing a brand note feature generation probability model according to the multi-peak normal distribution model; setting a weight value of each note index, adding each weight value into a brand note feature generation probability model, and constructing a similarity calculation formula; organizing smoke panel test personnel to perform smoke panel test on the tested cigarette, recording the score of each note index of the tested cigarette, substituting the score into the similarity calculation formula, and calculating a similarity score; and according to the similarity score, evaluating the similarity between the tested cigarette and the cigarette of the set brand. According to the method, the accuracy of the similarity measurement result of the cigarette note characteristics can be effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of cigarette evaluation, and in particular relates to a method for measuring the similarity of cigarette aroma characteristics. Background Art

[0002] In the process of cigarette production, in order to effectively test the degree of difference between different batches of cigarette products, evaluate the degree of difference between cigarette products processed in different places, examine the similarity between different brands, and evaluate the similarity between the aroma characteristics of new products and corporate brands, a similarity measurement method suitable for the evaluation of cigarette aroma characteristics is needed.

[0003] Compared with existing research results, the invention patent with publication number CN105335734A, "A method for identifying cigarette aroma characteristics by star graph", uses star graphs to compare aroma characteristics, but it cannot provide a specific score for similarity. The invention patent with publication number CN106770971A, "A method for evaluating cigarette style characteristics", uses contour graphs to analyze the evaluation results, but it cannot provide a specific score for similarity.

[0004] At present, the research on similarity measurement of cigarette aroma characteristics is mainly based on distance measurement. The "Comprehensive Evaluation of Cigarette Quality Stability-Based on Multi-Feature Similarity Analysis and Principal Component Analysis" of the Chinese Journal of Tobacco Science uses principal component analysis to reduce the dimension of cigarette feature data, and uses cosine similarity and Euclidean distance as similarity measurement. The "Application of Similarity Distance Analysis Method in Cigarette Sensory Evaluation Analysis" of the Journal of Jiangxi Agricultural Sciences summarizes the application of similarity distances such as Euclidean distance, Mahalanobis distance, and Minkowski distance in cigarette sensory evaluation methods. The "Establishment and Application of a Tobacco Sensory Evaluation Analysis Method" of the Journal of Tobacco Science and Technology proposed a sensory deviation evaluation index D value based on Euclidean distance. However, the similarity measurement method using Euclidean distance (i.e., total square difference or mean square error) has its potential problems. When using Euclidean distance for similarity measurement, the lower the total square difference between the test sample and the brand's core aroma characteristics in all aroma indicators, the more the product conforms to the brand characteristics. However, there are large differences between different aroma indicators. Some aroma indicators have large variances, the generated data is relatively discrete, and the corresponding distance fluctuations are large; some aroma indicators have large values, and the corresponding distances are often large; some aroma indicators can better highlight the brand characteristics, but they are not reflected in the traditional similarity measurement, resulting in the accuracy of their measurement results to be improved. Therefore, how to effectively improve the accuracy of the measurement results of the similarity of cigarette aroma characteristics has become a technical problem that technicians in this field urgently need to solve. Summary of the invention

[0005] The purpose of the present invention is to provide a method for measuring the similarity of cigarette aroma characteristics to solve the above technical problems in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for measuring similarity of cigarette aroma characteristics comprises the following steps:

[0008] Step S1, selecting representative cigarettes of different brands from a set brand of cigarettes, organizing smoking testers to smoke them, and recording the scores of various aroma indicators;

[0009] Step S2, calculating the mean and standard deviation of the scores of each flavor index in different brands of cigarettes, and constructing a multi-peak normal distribution model;

[0010] Step S3, constructing a brand fragrance feature generation probability model according to the multi-peak normal distribution model;

[0011] Step S4, setting weight values ​​for each fragrance index, and adding each weight value to the brand fragrance feature generation probability model to construct a similarity calculation formula;

[0012] Step S5, organizing smoking testers to smoke the test cigarettes, recording the scores of various aroma indicators of the test cigarettes, and substituting them into the similarity calculation formula to calculate the similarity score;

[0013] Step S6: Evaluate the similarity between the test cigarette and a set brand of cigarettes based on the similarity score.

[0014] Preferably, in step S3, the brand fragrance feature generation probability model is:

[0015]

[0016] Where: A=(a1,a2,a3,…,a k ), is the vector representation of the flavor characteristics of the test cigarette; a1, a2, a3, …, a k , is the vector representation of each flavor index of the test cigarette; θ=(θ1,θ2,θ3,…,θ k ), is the vector representation of the aroma characteristics of the set brand of cigarettes; θ1,θ2,θ3,…,θ k , which is the vector representation of each aroma index of the set brand of cigarettes.

[0017] Preferably, in step S4, the specific steps of setting the weight value of each fragrance index are:

[0018] Step S41, selecting other representative backbone brand cigarettes of the industry as industry samples representing the average level of the cigarette industry, organizing smoking testers to smoke them, and recording the scores of various aroma indicators;

[0019] Step S42: obtaining the test value of each flavor index of the set brand of cigarette relative to each flavor index of the industry sample through hypothesis testing, and using the test value as the weight value of the flavor index.

[0020] Preferably, in step S42, a one-tailed t-test is used as a hypothesis test, and the weight value of the aroma index = the logarithm of the t-value of the aroma index of a set brand of cigarettes divided by the sum of the logarithms of the t-values ​​of all aroma indexes of industry samples.

[0021] Preferably, in step S4, for each aroma index whose score distribution during inhalation evaluation conforms to a normal distribution, the similarity calculation formula is:

[0022]

[0023] Among them, S is the similarity; t j is the t value of the fragrance index of the industry sample; t i To set the t value of the aroma index of brand cigarettes, t i The value of is:

[0024]

[0025] Preferably, for each aroma index whose score distribution during smoking evaluation does not conform to the normal distribution, the lower limit value of S is set to 0, and the benchmark similarity when each aroma index of the test cigarette is equal to the average value of the set brand of cigarettes is calculated, recorded as S', and added as a subtrahend into the similarity calculation formula, that is:

[0026]

[0027] Preferably, it further comprises step S7: performing thermogravimetric analysis on the test cigarette and cigarettes of different brands that are representative of the set brand, and calculating the degree of pyrolysis difference.

[0028] Preferably, the method for obtaining the pyrolysis difference is:

[0029] The test cigarettes and each representative brand of the set brand were used as analysis samples. A certain amount of tobacco was taken and crushed using a high-speed grinder with a 60-80 mesh screen. 10-20 mg was weighed and placed in an alumina flat crucible. A thermal analyzer was used to heat the analysis sample from room temperature to 873K under N2 purge at a flow rate of 100 mL / min. The heating rate was 10K / min. At the same time, the data of thermogravimetry and differential scanning calorimetry were collected, and the thermal decomposition difference was calculated according to the following formula:

[0030]

[0031] Where: NRMSE represents the degree of pyrolysis difference; N represents the number of recording points in the thermogravimetric analysis in the range of 100-600°C; dm represents the mass percentage of the sample analyzed at time T; T represents time; The quality loss rate of the lth recording point of the set brand; It is the mass loss rate of the lth recording point of the test cigarette.

[0032] The beneficial effects of the present invention are:

[0033] The method for measuring the similarity of the aroma characteristics of cigarettes of the present invention constructs a probability model for generating brand aroma characteristics, and sets the weight value of each aroma index, thereby constructing a similarity calculation formula, and then substitutes the score of each aroma index of the test cigarette during the evaluation into the similarity calculation formula, calculates the similarity score, and can conveniently judge the similarity between the aroma characteristics of the test cigarette and the set brand cigarette according to the similarity score. It can be seen that the similarity calculation formula of the present invention not only utilizes the probability model for generating brand aroma characteristics, but also sets a weight value for each aroma index, effectively avoiding the problems caused by using Euclidean distance for similarity measurement in the prior art, thereby effectively improving the accuracy of the measurement results of the similarity of the aroma characteristics of cigarettes. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the embodiments, and further describe the specific embodiments of the present invention in detail in conjunction with the drawings, wherein

[0035] Figure 1 A flowchart of a method for measuring similarity of cigarette aroma characteristics provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the technical solution of the present invention, the solution will be further described in detail below in conjunction with specific embodiments.

[0037] like Figure 1 As shown, the embodiment of the present invention provides a method for measuring the similarity of cigarette aroma characteristics, which includes the following steps:

[0038] Step S1, selecting representative cigarettes of different brands from a set brand of cigarettes, organizing smoking testers to smoke them, and recording the scores of various aroma indicators;

[0039] Step S2, calculating the mean and standard deviation of the scores of each flavor index in different brands of cigarettes, and constructing a multi-peak normal distribution model;

[0040] Step S3, constructing a brand fragrance feature generation probability model according to the multi-peak normal distribution model;

[0041] Step S4, setting weight values ​​for each fragrance index, and adding each weight value to the brand fragrance feature generation probability model to construct a similarity calculation formula;

[0042] Step S5, organizing smoking testers to smoke the test cigarettes, recording the scores of various aroma indicators of the test cigarettes, and substituting them into the similarity calculation formula to calculate the similarity score;

[0043] Step S6: Evaluate the similarity between the test cigarette and a set brand of cigarettes based on the similarity score.

[0044] The method for measuring the similarity of the aroma characteristics of cigarettes provided in the embodiment of the present invention constructs a probability model for generating brand aroma characteristics, and sets the weight value of each aroma index, thereby constructing a similarity calculation formula, and then substitutes the score of each aroma index of the test cigarette during the evaluation into the similarity calculation formula, calculates the similarity score, and can conveniently judge the similarity between the aroma characteristics of the test cigarette and the set brand cigarette according to the similarity score. It can be seen that the similarity calculation formula of the present invention not only utilizes the probability model for generating brand aroma characteristics, but also sets a weight value for each aroma index, effectively avoiding the problems caused by using Euclidean distance for similarity measurement in the prior art, thereby effectively improving the accuracy of the measurement results of the similarity of the aroma characteristics of cigarettes.

[0045] Specifically, YC / T497-2014 “Sensory Evaluation Method for Chinese Cigarette Style” can be used to evaluate and score various aroma indicators.

[0046] Furthermore, in step S3, the probability model for generating brand fragrance features is:

[0047]

[0048] Where: A=(a1,a2,a3,…,a k ), is the vector representation of the flavor characteristics of the test cigarette; a1, a2, a3, …, a k , is the vector representation of each flavor index of the test cigarette; θ=(θ1,θ2,θ3,…,θ k ), is the vector representation of the aroma characteristics of the set brand of cigarettes; θ1,θ2,θ3,…,θ k , which is the vector representation of each aroma index of the set brand of cigarettes.

[0049] It should be noted that the probability model for generating brand fragrance features can originally be:

[0050]

[0051] However, in order to facilitate calculation and avoid numerical underflow during the implementation of the algorithm, a negative logarithmic function is used to transform the generated probability model, which is converted into the above formula (1-1).

[0052] It is understandable that the smaller the value of the negative logarithm of the generation probability, the more inclined it is to the aroma characteristics of the set brand. However, if only P(A|Θ) is considered, the following situations that affect the judgment may occur: for example, there are two test cigarettes A1 and A2, P(A1|Θ) is slightly greater than P(A2|Θ), but the test cigarette A2 performs more similarly in the more prominent aroma indicators of the set brand. At this time, if only P(A|Θ) is compared, it may be concluded that the test cigarette A1 is more in line with the aroma characteristics of the brand, but considering P(A|Θ) and its performance in the more prominent aroma indicators of the set brand, the real conclusion may be that the test cigarette A2 is closer to the aroma characteristics of the set brand. In order to avoid the above problems, it is necessary to set weight values ​​for each aroma indicator.

[0053] In step S3, when constructing a probability model for generating brand aroma characteristics according to a multi-peak normal distribution model, since the probability density function of the normal distribution is continuous, during the implementation of this embodiment, the value interval of the normal distribution (μ-3*σ, μ+3*σ) can be evenly divided into multiple small intervals with a unit length of 0.01, and 1,000,000 random data are generated according to the Box-Muller algorithm, and the probability of falling in each small interval is calculated as a method of discretizing a continuous function. Where μ is the mean and σ is the standard deviation.

[0054] In order to make the similarity measurement highlight the aroma characteristics of the set brand that are of high concern, it is necessary to evaluate the degree of difference of the set brand in each aroma index relative to the overall level of the industry, and use it as the weight value of each aroma index when measuring the similarity. Taking flue-cured tobacco cigarettes as an example, since there are hundreds of different brands of flue-cured tobacco cigarette products on the market, it can be considered that the score distribution of flue-cured tobacco cigarettes in each aroma index conforms to the normal distribution. The normal distribution of aroma indicators provides a prerequisite for hypothesis testing. The test value obtained through hypothesis testing reflects the difference between the aroma indicators of the brand and other brands in the industry, and can be used to indicate the degree to which each aroma indicator highlights the aroma characteristics of the brand. Therefore, in step S4, the specific steps for setting the weight value of each aroma indicator are:

[0055] Step S41, selecting other representative backbone brand cigarettes of the industry as industry samples representing the average level of the cigarette industry, organizing smoking testers to smoke them, and recording the scores of various aroma indicators;

[0056] Step S42: obtaining the test value of each flavor index of the set brand of cigarette relative to each flavor index of the industry sample through hypothesis testing, and using the test value as the weight value of the flavor index.

[0057] Furthermore, in step S42, a one-tailed t-test is used as a hypothesis test, and the weight value of the aroma index = the logarithm of the t-value of the aroma index of the set brand of cigarettes divided by the sum of the logarithms of the t-values ​​of all aroma indexes of industry samples.

[0058] Specifically, in step S4, for each aroma index whose score distribution during inhalation evaluation conforms to the normal distribution, the similarity calculation formula is:

[0059]

[0060] Among them, S is the similarity; t j is the t value of the jth fragrance index of the industry sample; t i To set the t value of the i-th aroma index of brand cigarettes, i.e., the test value, t i The value of is:

[0061]

[0062] It should be noted that due to t i The absolute value of t may be less than 1, so it is necessary to i Do the corresponding processing: when t i When the absolute value of is less than or equal to 1, it means that the set brand of cigarettes is almost the same as other brands of cigarettes in the industry in terms of the aroma index. At this time, the numerator of the weight item is set to 0, even if t i =1, indicating that the fragrance index contributes little to the fragrance characteristics of the brand; when t i When the absolute value of is greater than 1, t i The value of is not processed. i The absolute value of is large, indicating that the set brand of cigarettes is significantly different from other brands of cigarettes in the industry in terms of the aroma index. At this time, the weight is large, and the contribution of its generation probability to the similarity S will be much greater than that of other aroma indicators.

[0063] When the score distribution of certain aroma indexes does not conform to the normal distribution, such as the absence of "air-dried tobacco aroma" in most flue-cured tobacco cigarettes, "green fragrance" is mainly reflected in "Nanjing" cigarettes, and "herbal fragrance" is mainly reflected in "Jinsheng" cigarettes, then it is necessary to add correction items. When the generation probability P(A|Θ) is larger, S is smaller, and when S is smaller, the test cigarette A is closer to the aroma characteristics of the set brand of cigarettes. Therefore, for the convenience of representation, the lower limit value of S is set to 0, and the baseline similarity when the average value of the aroma index of the test cigarette is equal to the aroma index of the set brand of cigarettes is calculated, recorded as S', and added as a subtrahend to the similarity calculation formula, that is:

[0064]

[0065] It is understood that for each fragrance index whose score distribution during the evaluation of the inhalation conforms to the normal distribution, the formula (2-1) is used to calculate the similarity. For each fragrance index whose score distribution during the evaluation of the inhalation does not conform to the normal distribution, the formula (2-2) is used to calculate the similarity.

[0066] In other embodiments, when the amount of data is sufficient, a one-tailed z test may also be used as a hypothesis test, which will not be described in detail here.

[0067] Furthermore, the method for measuring the similarity of cigarette aroma characteristics also includes step S7: performing thermogravimetric analysis on the test cigarette and cigarettes of different brands that are representative of the set brand, and calculating the pyrolysis difference to verify the effectiveness and feasibility of the measurement method.

[0068] Specifically, the method for obtaining the pyrolysis difference is:

[0069] The test cigarettes and each representative brand of the set brand are used as analysis samples. A certain amount of tobacco is taken and crushed using a high-speed crusher with a 60-80 mesh screen. 10-20 mg is weighed and placed in an alumina flat crucible. A thermal analyzer is used to heat the analysis sample from room temperature to 873K, i.e., 600°C, with a heating rate of 10K / min under N2 purge at a flow rate of 100mL / min. At the same time, the data of thermogravimetry and differential scanning calorimetry are collected, and the thermal decomposition difference is calculated according to the following formula:

[0070]

[0071] Where: NRMSE represents the degree of pyrolysis difference; N represents the number of recording points in the thermogravimetric analysis range of 100-600℃; dm represents the mass percentage of the sample analyzed at time T, in units of %; T represents time, in units of min; The mass loss rate of the first recording point of the set brand, unit: % / min; It is the mass loss rate of the lth recording point of the tested cigarette, in % / min.

[0072] It should be noted that for cigarette companies, cigarette products of the same brand often have similar aroma characteristics, such as the herbal aroma of "Jinsheng" cigarettes, the green aroma of "Nanjing" cigarettes, and the flue-cured tobacco aroma of "Golden Leaf" cigarettes. Under the influence of market competition and corporate innovation and development, the products of various cigarette companies have gradually formed brand-specific aroma characteristics. Therefore, measuring the similarity between the test sample and the brand aroma characteristics is of great significance to product design, product development, market analysis and other work.

[0073] The present invention selects representative cigarettes of different brands under the same brand to form a reference sample, and the generation probability of the test sample under the reference sample generation model can be used to measure the similarity between the test sample and the reference sample. In the production process, due to the fluctuation of tobacco leaf quality, the fluctuation of flavoring, adding materials, and tobacco blending, the fluctuation of production, transportation, storage environment and other factors, the aroma characteristics of cigarettes will change within a certain range. Therefore, for a single cigarette in the reference sample, it can be considered that the distribution of its aroma index score conforms to the normal distribution, and the distribution of the entire reference sample conforms to the multi-peak normal distribution. According to the distribution of the aroma characteristics of the reference sample cigarettes, a generation probability model of the test sample can be established, and the level of the generation probability is used as a metric.

[0074] Comparative test:

[0075] Table 1 shows the similarity measurement of four industry backbone brands of cigarettes and four different brands of "Golden Leaf" cigarettes as test samples using Euclidean distance, cosine similarity, and "Golden Leaf" brand aroma feature similarity measurement method (i.e., the above-mentioned set brand of cigarettes is selected as "Golden Leaf") respectively.

[0076] Table 1 Similarity measurement results and pyrolysis difference of different measurement methods

[0077]

[0078] From Table 1 we can see that:

[0079] The Euclidean distance between the "Golden Leaf" brand cigarette test sample and the "Golden Leaf" cigarette average level is relatively low, indicating that the distance between the "Golden Leaf" brand cigarette test sample and the "Golden Leaf" average level is closer. However, the scores of "Core Brand Sample 2" and "Golden Leaf Sample 3" are at the same level, indicating that using the Euclidean distance as a similarity measure fails to consider the distribution of each aroma index, and thus an inaccurate measurement result is obtained;

[0080] Through the cosine similarity, we can see that the scores of the "Golden Leaf" brand test samples are generally higher than other backbone brands in the industry. The vectors representing the evaluation results of the "Golden Leaf" brand cigarette test samples have a smaller angle with the vectors formed by the average level of "Golden Leaf" cigarettes, and the similarity is higher. However, the results of the cosine similarity are not obvious. The difference can only be seen when retaining 4 decimal places. The difference is very small and inaccurate;

[0081] The similarity S shows that the score of the "Golden Leaf" brand cigarette test sample is significantly lower than that of other industry backbone brands. According to the principle that the lower the score, the higher the similarity, it can be concluded that the "Golden Leaf" brand test sample is more similar to the "Golden Leaf" brand fragrance characteristics, which is in line with experimental expectations.

[0082] It can be seen from the pyrolysis difference values ​​that the pyrolysis difference between the "Golden Leaf" brand and the "Golden Leaf" test sample is also small, which verifies the feasibility of the similarity measurement method of the present invention.

[0083] The method for measuring the similarity of cigarette aroma characteristics proposed in the present invention overcomes the defects of the prior art, does not have problems such as classification boundary crossing and low recognition, can effectively improve the recognition of product maintenance results by product personnel, and has guiding significance for new products. At the same time, in order to verify the effectiveness and feasibility of using the generative probability model as a measurement method for the similarity of cigarettes, the present invention adopts a thermogravimetric-based analysis method, sets brand cigarettes and other backbone brand cigarettes representing the overall level of the industry as analysis samples, uses the same thermal analysis method to conduct thermogravimetric experiments, and uses the standard square root error to evaluate the thermal decomposition difference between the analysis samples to verify the effectiveness and feasibility of the above measurement method. The present invention proposes a method for measuring the similarity of cigarette aroma characteristics combined with thermogravimetric analysis, which can be used to measure the similarity or difference between the test sample and the aroma characteristics of a single brand of cigarettes or brand.

[0084] The above are only preferred embodiments of the present invention. It should be pointed out that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. Moreover, after reading the contents of the present invention, relevant technical personnel in the field may make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the claims attached to this application.

Claims

1. A method for measuring the similarity of cigarette aroma characteristics, characterized in that: It includes the following steps: Step S1, selecting representative cigarettes of different brands from a set brand of cigarettes, organizing smoking testers to smoke them, and recording the scores of various aroma indicators; Step S2, calculating the mean and standard deviation of the scores of each flavor index in different brands of cigarettes, and constructing a multi-peak normal distribution model; Step S3, constructing a brand fragrance feature generation probability model according to the multi-peak normal distribution model; Step S4, setting weight values ​​for each fragrance index, and adding each weight value to the brand fragrance feature generation probability model to construct a similarity calculation formula; Step S5, organizing smoking testers to smoke the test cigarettes, recording the scores of various aroma indicators of the test cigarettes, and substituting them into the similarity calculation formula to calculate the similarity score; Step S6: Evaluate the similarity between the test cigarette and a set brand of cigarettes based on the similarity score.

2. The method for measuring similarity of cigarette aroma characteristics according to claim 1, characterized in that: In step S3, the probability model for generating brand fragrance features is: Where: A=(a1,a2,a3,…,a k ), is the vector representation of the flavor characteristics of the test cigarette; a1, a2, a3, …, a k , is the vector representation of each flavor index of the test cigarette; θ=(θ1,θ2,θ3,…,θ k ), is the vector representation of the aroma characteristics of the set brand of cigarettes; θ1,θ2,θ3,…,θ k , which is the vector representation of each aroma index of the set brand of cigarettes.

3. The method for measuring similarity of cigarette aroma characteristics according to claim 2, characterized in that: In step S4, the specific steps of setting the weight value of each fragrance index are: Step S41, selecting other representative backbone brand cigarettes of the industry as industry samples representing the average level of the cigarette industry, organizing smoking testers to smoke them, and recording the scores of various aroma indicators; Step S42: obtaining the test value of each flavor index of the set brand of cigarette relative to each flavor index of the industry sample through hypothesis testing, and using the test value as the weight value of the flavor index.

4. The method for measuring similarity of cigarette aroma characteristics according to claim 3, characterized in that: In step S42, a one-tailed t-test is used as a hypothesis test, and the weight value of the aroma index = the logarithm of the t-value of the aroma index of a set brand of cigarettes divided by the sum of the logarithms of the t-values ​​of all aroma indexes of industry samples.

5. The method for measuring similarity of cigarette aroma characteristics according to claim 4, characterized in that: In step S4, for each aroma index whose score distribution during the evaluation conforms to the normal distribution, the similarity calculation formula is: Among them, S is the similarity; t j is the t value of the fragrance index of the industry sample; t i To set the t value of the aroma index of brand cigarettes, t i The value of is:

6. The method for measuring similarity of cigarette aroma characteristics according to claim 4, characterized in that: For each aroma index whose score distribution during smoking evaluation does not conform to the normal distribution, the lower limit value of S is set to 0, and the baseline similarity when the aroma index of the test cigarette is equal to the average value of the aroma index score of the set brand of cigarettes is calculated, recorded as S', and added as a subtrahend into the similarity calculation formula, that is:

7. The method for measuring similarity of cigarette aroma characteristics according to any one of claims 1 to 6, characterized in that: The method further comprises step S7: performing thermogravimetric analysis on the test cigarette and cigarettes of different brands that are representative of the set brand, and calculating the degree of pyrolysis difference.

8. The method for measuring similarity of cigarette aroma characteristics according to claim 7, characterized in that: The method for obtaining the pyrolysis difference is: The test cigarettes and each representative brand of the set brand were used as analysis samples. A certain amount of tobacco was taken and crushed using a high-speed grinder with a 60-80 mesh screen. 10-20 mg was weighed and placed in an alumina flat crucible. A thermal analyzer was used to heat the analysis sample from room temperature to 873K under N2 purge at a flow rate of 100 mL / min. The heating rate was 10K / min. At the same time, the data of thermogravimetry and differential scanning calorimetry were collected, and the thermal decomposition difference was calculated according to the following formula: Where: NRMSE represents the degree of pyrolysis difference; N represents the number of recording points in the thermogravimetric analysis in the range of 100-600°C; dm represents the mass percentage of the sample analyzed at time T; T represents time; The quality loss rate of the lth recording point of the set brand; It is the mass loss rate of the lth recording point of the test cigarette.

Images