Product quality stability evaluation method based on entropy weight method

Abstract

The invention provides a product quality stability evaluation method based on an entropy weight method, which comprises the following steps of: scanning cut tobacco spectrums of cigarette products which have the same mark and are produced in different places and at different times by using a near-infrared spectrometer, and predicting chemical components by using a tobacco near-infrared big data system; the key chemical components of the tobacco shreds are assigned through an entropy weight method, a quality stability model is constructed by using the assigned indexes, the model is composed of a mean value after entropy weight and an identification interval, and the model constructed by the method is high in identification accuracy, high in detection efficiency, low in cost, green, environmentally friendly and easy to operate. And a technical guarantee is provided for product stability analysis of cigarette enterprises.

Description

Technical Field

[0001] The invention belongs to the technical field of near-infrared nondestructive testing, and specifically relates to a product quality stability evaluation method based on an entropy weight method, which is used for quickly analyzing the quality stability of cigarette products. Background Art

[0002] Tobacco quality stability is an important indicator of cigarette product quality. For a long time, the intrinsic quality of cigarette products has been mainly characterized by sensory evaluation, lacking an intuitive and systematic quantitative description method. The quality stability evaluation of cigarette products mainly uses continuous flow method, HPLC and GC-MS chromatography to detect trace chemical components in tobacco for stability evaluation. Although a large amount of information on the components contained in cigarette samples can be obtained through separation and analysis of chromatographic instruments, the detection process is slow and the correlation analysis of data is complicated due to limitations in sample preparation, test conditions, time, etc. By the time the analysis results come out, the product has already entered the market and cannot be improved. At present, the competition in the industry product market is becoming increasingly fierce, and there is an urgent need for fast and convenient methods to improve product quality.

[0003] Near-infrared spectroscopy is a green and environmentally friendly rapid analysis technology. A near-infrared spectrum contains rich physical and chemical information of the sample. By combining the multivariate analysis methods in modern chemometrics (such as multivariate calibration and pattern recognition) with near-infrared spectroscopy analysis technology and establishing corresponding calibration models and classification models, complex multi-component systems can be analyzed. This can not only predict multiple components or properties of unknown samples, but also classify their quality attributes.

[0004] This patent establishes a product quality stability evaluation method based on the entropy weight method. According to the discreteness of the data itself, the entropy value of each indicator is calculated to evaluate the amount of information provided by it, so as to determine the weight. It mainly includes the following steps: first, the tobacco of cigarette products is scanned by a near-infrared spectrometer; then the tobacco near-infrared big data application system is used to predict the chemical composition; the eight indicator parameters of the key chemical components of tobacco, such as total sugar, total nitrogen, and total alkaloids, are assigned values ​​by the entropy weight method, and models are constructed for each indicator of the key chemical components, providing a scientific basis for the stability analysis of quality products. Summary of the invention

[0005] The purpose of the present invention is to establish a product quality stability evaluation method based on the entropy weight method, determine the product stability fluctuation threshold, and realize rapid evaluation of the quality stability of cigarette products.

[0006] The present invention uses a near-infrared spectrometer to grind a series of tobacco shreds of the same brand, different batches of products at different times into tobacco powder for near-infrared scanning, and uses a tobacco near-infrared big data system to predict chemical components; assigns values ​​to key chemical components of tobacco shreds by the entropy weight method; and uses the assigned indicators to construct a model, which consists of the mean value and recognition interval after entropy weighting. The model built by this method has high recognition accuracy, high detection efficiency, low cost, and is green, environmentally friendly, and easy to operate, providing technical support for product stability analysis of cigarette companies.

[0007] A product quality stability evaluation method based on entropy weight method comprises the following steps: step 1) scanning tobacco of cigarette products by near infrared spectrometer; step 2) using tobacco near infrared big data system to predict chemical composition; step 3) assigning values ​​to key chemical components of tobacco by entropy weight method; step 4) constructing models for various indicators of key chemical components and determining product quality stability identification interval.

[0008] The specific modeling steps are as follows:

[0009] (1) For n samples and m indicators, then X ij is the value of the jth indicator of the ith sample (i=1,…,n; j=1,…,m).

[0010] (2) Normalization of key indicators: Key indicators are divided into positive and negative indicators, and different algorithms are required for data normalization. The values ​​of positive and negative indicators have different meanings (the higher the value of the positive indicator, the better; the lower the value of the negative indicator, the better). After normalization, the absolute value of the indicator is converted into a relative value, thereby solving the stability problem of the values ​​of different indicators.

[0011] 1) Positive indicators (the larger the value, the better):

[0012]

[0013] 2) Negative indicators (the smaller the value, the better):

[0014]

[0015] Get the standardized data set x ij

[0016] (3) Calculate the proportion: Calculate the proportion of the i-th sample value under the j-th indicator to the indicator:

[0017]

[0018] (4) Calculate the entropy value: Calculate the entropy value of the jth indicator:

[0019]

[0020] in, Meet e j ≥0. The larger the entropy value, the smaller the amount of effective information provided by the indicator, and the smaller its weight; conversely, the smaller the entropy value, the greater the amount of effective information provided by the indicator, and the greater its weight.

[0021] (5) Calculate information entropy redundancy (difference)

[0022] d j =1-e j , j = 1, ..., m

[0023] (6) Calculate weight: Calculate the weight of the jth indicator:

[0024]

[0025] The weights w of the eight indicators are obtained through the above calculations. j , we can make a comprehensive evaluation of each object.

[0026] (7) Calculate the comprehensive score of each batch of finished tobacco samples:

[0027]

[0028] Among them, x ij is the standardized data.

[0029] (8) Calculate the average comprehensive score of each batch of finished tobacco:

[0030]

[0031] (9) According to the accuracy requirements of product stability, set the stability range of the comprehensive score of finished tobacco:

[0032]

[0033] The value of Δs can be set manually based on the actual scores of finished tobacco of different specifications of cigarettes and the requirements for stability control accuracy. The quality of the i-th batch is stable if Then the i-th batch is considered to have quality deviation.

[0034] The steps before scanning include: crushing the sample into 40-80 mesh, and the sample is tobacco shreds, tobacco stems and / or tobacco dust.

[0035] The beneficial effects of the present invention are as follows:

[0036] 1. The present invention uses a PCA algorithm to perform qualitative analysis on near-infrared spectra and establish a similarity classification and recognition model for tobacco leaf producing areas in different regions.

[0037] 2. The method of the present invention is simple, quick, non-destructive to samples and non-polluting to the environment.

[0038] 3. The present invention has the advantages of low cost and high efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is the original near-infrared scanning spectrum of cigarette products;

[0040] Figure 2 It is the quality stability model of cigarette brand “A” based on entropy weight method;

[0041] Figure 3 It is the quality stability model of cigarette brand “A” based on entropy weight method; DETAILED DESCRIPTION

[0042] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0043] Example 1

[0044] 1. Experimental instruments

[0045] MPA Fourier near-infrared spectrometer, ZM200 precision pulverizer, and oven produced by BRUKER (Germany).

[0046] 2. Sample collection

[0047] In order to make the established classification model have wide applicability, a total of 30 samples of "A" brand products of the same brand produced by Guangxi Zhongyan Nanning Factory in different time periods were selected for modeling.

[0048] 3. Sample preparation

[0049] The cigarette shreds were dried in an oven at 40°C to keep the moisture content of the samples basically consistent, and then fully crushed with a ZM200 precision grinder, passed through a 60-mesh sieve, and adjusted to a moisture content of 6-8%.

[0050] 4. Spectral scanning and data processing

[0051] The spectrum of tobacco leaf samples was scanned using an MPA Fourier near-infrared spectrometer (with a near-infrared quantitative analysis diffuse reflectance gold-plated large integrating sphere and sample rotator sampling accessories) produced by BRUKER (Germany), and the spectrum was processed using the qualitative analysis software QUANT7.0 in Bruker OPUS. The specific operation is as follows: put tobacco powder into the sample cup, the height in the cup is about 3 cm, press the weight on the sample for 10 seconds and then take it out, wipe the quartz glass at the bottom of the cup with gauze, and then place the sample cup on a rotating platform for NIR scanning. The operating parameters are: spectral scanning range 12 000~4 000cm -1 , spectral resolution 8cm -1 , scanning times 64 times (about 30S). Spectral data were collected in a transmission mode and processed into the first-order differential of the absorption spectrum. The original scan of the cigarette tobacco is shown in Figure 1 In the modeling process, in order to eliminate the influence of noise and baseline, the standard normal variable transformation method was used to eliminate the differences caused by the particle size, surface scattering and optical path change of solid samples, the spectrum was smoothed and filtered by wavelet transform, and the second-order derivative was used for preprocessing. After the sample was scanned, the spectral data was processed using statistical software.

[0052] 5. Using the “Tobacco Near Infrared Big Data Application System” to predict the chemical composition of tobacco

[0053] The predicted results of the main chemical compositions of 30 batches of "A" brand products produced by Guangxi China Tobacco Co., Ltd. in 2023 are shown in Table 1

[0054] Table 1 Prediction results of “Tobacco Near Infrared Big Data Application System” (excerpt of main components)

[0055]

[0056]

[0057] 6. Establishment of the entropy weight method model

[0058] The steps to build the model are as follows:

[0059] (1) For n samples and m indicators, then X ij is the value of the jth indicator of the ith sample (i=1, ..., n; j=1, ..., m).

[0060] (2) Normalization of key indicators: Key indicators are divided into positive and negative indicators, and different algorithms are required for data normalization. The values ​​of positive indicators and negative indicators have different meanings (the higher the value of the positive indicator, the better; the lower the value of the negative indicator, the better). After normalization, the absolute value of the indicator is converted into a relative value, thereby solving the stability problem of the values ​​of different indicators.

[0061] 1) Positive indicators (the larger the value, the better):

[0062]

[0063] 2) Negative indicators (the smaller the value, the better):

[0064]

[0065] Get the standardized data set x ij

[0066] (3) Calculate the proportion: Calculate the proportion of the i-th sample value under the j-th indicator to the indicator:

[0067]

[0068] (4) Calculate the entropy value: Calculate the entropy value of the jth indicator:

[0069]

[0070] in, Meet e j ≥0. The larger the entropy value, the smaller the amount of effective information provided by the indicator, and the smaller its weight; conversely, the smaller the entropy value, the greater the amount of effective information provided by the indicator, and the greater its weight.

[0071] (5) Calculate information entropy redundancy (difference)

[0072] d j =1-e j , j = 1, ..., m

[0073] (6) Calculate weight: Calculate the weight of the jth indicator:

[0074]

[0075] The weights w of the eight indicators are obtained through the above calculations. j , we can make a comprehensive evaluation of each object.

[0076] (7) Calculate the comprehensive score of each batch of finished tobacco samples:

[0077]

[0078] Among them, x ij is the standardized data.

[0079] (8) Calculate the average comprehensive score of each batch of finished tobacco:

[0080]

[0081] (9) According to the accuracy requirements of product stability, set the stability range of the comprehensive score of finished tobacco:

[0082] The value of Δs can be set manually based on the actual scores of finished tobacco of different specifications of cigarettes and the requirements for stability control accuracy. The quality of the i-th batch is stable if Then the i-th batch is considered to have quality deviation.

[0083] According to the data in Table 1, the weights of 8 chemical indicators were calculated using the entropy weight method: [0.07811578, 0.12363959, 0.14781896, 0.14375827, 0.14939766, 0.11505525, 0.13050942, 0.11170507].

[0084] The comprehensive evaluation method was used to calculate the comprehensive scores of the test data of 30 batches of "A" brand finished cigarettes. The entropy weight method was used to weight each indicator, and the entropy values ​​of the eight chemical indicators were calculated as follows: [0.96759819, 0.94871526, 0.93868585, 0.94037019, 0.93803102, 0.95227598, 0.94586571, 0.95366561]. Table 2: Scores of 30 batches of "A" brand cigarettes:

[0085] Table 2 Scores of 30 batches of “A” brand cigarettes

[0086] Batch number Comprehensive score Batch number Comprehensive score 1 8.523492748 16 8.480448983 2 8.406239882 17 8.442110162 3 8.333587440 18 8.444973663 4 8.556083677 19 8.414517431 5 8.521068432 20 8.559268128 6 8.292799179 21 8.362268054 7 8.311283092 22 8.275164502 8 8.744265653 23 8.433554352 9 8.541007816 24 8.534581801 10 8.530755531 25 8.517396944 11 8.770283146 26 8.532864138 12 8.626561263 27 8.564855891 13 8.503401652 28 8.408067187 14 8.496851139 29 8.412712816 15 8.537949067 30 8.329081147

[0087] The mean (score_mean) of the comprehensive scores of 30 batches of samples of brand “A” is: 8.48024983.

[0088] After comprehensive analysis, the lowest comprehensive score of the 30 batches of "A" is 8.275164501856096, the highest is 8.770283146129488, and the average score is: 8.48024983. The scatter plot of the comprehensive scores of the 30 batches is as follows, where the horizontal axis is the batch number and the vertical axis is the comprehensive score. The stability interval of the "A" product is determined to be [score_mean-0.25, score_mean+0.25]. That is, the comprehensive score in the scatter plot is within the two red line interval to meet the stability requirements, and a stability warning will be issued if it exceeds the two red line interval. Figure 2 .

[0089] From the data results and scatter plot, we can see that batches 8 and 11 of brand “A” cigarettes are beyond the stability area, indicating that these two batches of products have fluctuations and need special attention.

[0090] The quality stability model of brand “A” cigarette products is established in Figure 2 :

[0091] Example 2

[0092] In this embodiment, the model is applied to the homogenization analysis of the same brand of cigarettes "B" produced in different places of origin.

[0093] 1. Experimental instruments

[0094] MPA Fourier near-infrared spectrometer, ZM200 precision pulverizer, and oven produced by BRUKER (Germany).

[0095] 2. Sample collection

[0096] In order to make the established classification model have wide applicability, a total of 64 samples of "B" brand products of the same brand produced by Guangxi Tobacco Nanning and Liuzhou factories in different time periods and different production lines were selected for modeling.

[0097] 3. Sample preparation

[0098] The cigarette shreds were dried in an oven at 40°C to keep the moisture content of the samples basically consistent, and then fully crushed with a ZM200 precision grinder, passed through a 60-mesh sieve, and adjusted to a moisture content of 6-8%.

[0099] 4. Spectral scanning and data processing

[0100] Spectral scanning and data processing are the same as in Example 1.

[0101] 5. Using the “Tobacco Near Infrared Big Data Application System” to predict the chemical composition of tobacco

[0102] The predicted results of the main chemical compositions of 64 batches of "B" brand products produced by Guangxi China Tobacco Co., Ltd. in Nanning and Liuzhou in 2023 are shown in Table 2

[0103] Table 3 Prediction results of “Tobacco Near Infrared Big Data Application System” (excerpt of main components)

[0104]

[0105]

[0106] 6. Establishment of the entropy weight method model

[0107] The steps of establishing the model are the same as those in Example 1 above.

[0108] According to the data in Table 1, the weights of 8 chemical indicators were calculated using the entropy weight method: [0.0648443, 0.15571385, 0.13768633, 0.14638612, 0.09875376, 0.06958399, 0.16363549, 0.16339615].

[0109] The comprehensive evaluation method was used to calculate the comprehensive scores of the test data of 64 batches of finished cigarettes of brand "B". The entropy weight method was used to weight each indicator, and the entropy values ​​of 8 chemical indicators were calculated to be: [0.98637747, 0.96728754, 0.97107477, 0.96924711, 0.97925375, 0.98538175, 0.96562336, 0.96567364].

[0110] Table 4 Scores of 64 batches of “A” brand cigarettes

[0111]

[0112]

[0113] The mean (score_mean) of the comprehensive scores of 64 batches of samples of brand “B” is: 7.71394838.

[0114] After comprehensive analysis, the lowest comprehensive score of the 64 batches of "B" was 7.3555903037059025, the highest was 8.537114444524637, and the average score was 7.71394838. The scatter plot of the comprehensive scores of the 64 batches is as follows, where the horizontal axis is the batch number and the vertical axis is the comprehensive score. The stability range of the "B" product is determined to be [score_mean-0.28, score_mean+0.28]. That is, if the comprehensive score in the scatter plot is within the two red line range, the stability requirements are met, and if it exceeds the two red line range, a stability warning will be issued.

[0115] from Figure 3 From the data results and scatter plot, we can see that batches 23, 24, 37, and 46 of "B" brand cigarettes are beyond the stability zone, with batches 37 and 46 being farther away, indicating that the products of these two batches are more volatile and require special attention.

[0116] The quality stability model of "B" brand cigarette products is established in Figure 3 :

[0117] 7. Model Validation

[0118] In order to better verify the recognition ability of the model, this experiment adopts the method of evaluation and verification. Professional evaluation experts conduct secret evaluation on samples that the model identifies as fluctuating. The results are shown in Table 3:

[0119] Table 5 Evaluation and verification recognition results

[0120]

[0121] The results showed that the eight samples from different batches and origins that were identified as having quality fluctuations by the model were all verified by smoking and found to have sensory fluctuations, with an identification rate of 100%, indicating that the established model has high prediction accuracy and can be used for quality stability analysis of cigarette products.

[0122] It can be seen from the above embodiments that the present invention utilizes a product quality stability evaluation model based on the entropy weight method, which has a high recognition rate for quality fluctuations of cigarettes. It can be seen that the application of near-infrared analysis technology can be well used for cigarette quality stability analysis, which is an effective and feasible method with strong practicality and realistic significance.

[0123] The above embodiments of the present invention are merely examples to clearly illustrate the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to enumerate the implementation methods used here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A product quality stability evaluation method based on entropy weight method, characterized in that: The following steps are involved: Step 1) scanning the tobacco of the cigarette product by a near infrared spectrometer; Step 2) predicting the chemical composition by using a tobacco near infrared big data system; Step 3), assigning values ​​to the key chemical components of tobacco by using the entropy weight method; Step 4), constructing models for various indicators of key chemical components and determining the product quality stability identification range.

2. According to claim 1, a method for evaluating product quality stability based on entropy weight method is characterized in that: The model construction includes the following steps: (1) For n samples and m indicators, then X ij is the value of the jth indicator of the i-th sample (i=1,…,n; j=1,…,m), (2) Normalization of key indicators: Key indicators are divided into positive and negative indicators, and different algorithms are required for data normalization. The values ​​of positive indicators and negative indicators have different meanings (the higher the value of positive indicators, the better; the lower the value of negative indicators, the better). After normalization, the absolute value of the indicator is converted into a relative value, thereby solving the stability problem of the values ​​of different indicators. 1) Positive indicators (the larger the value, the better): 2) Negative indicators (the smaller the value, the better): Get the standardized data set x ij (3) Calculate the proportion: Calculate the proportion of the i-th sample value under the j-th indicator to the indicator: (4) Calculate the entropy value: Calculate the entropy value of the jth indicator: in, Meet e j ≥0; the larger the entropy value, the smaller the effective information provided by the indicator, and the smaller its weight; conversely, the smaller the entropy value, the greater the effective information provided by the indicator, and the greater its weight; (5) Calculate information entropy redundancy (difference) d j =1-e j ,j=1,…,m (6) Calculate weight: Calculate the weight of the jth indicator: The weights w of the eight indicators are obtained through the above calculations. j , we can make a comprehensive evaluation of each object; (7) Calculate the comprehensive score of each batch of finished tobacco samples: Among them, x ij is the standardized data; (8) Calculate the average comprehensive score of each batch of finished tobacco: (9) According to the accuracy requirements of cigarette product stability, the stability range is set as: The value of Δs can be set manually according to the actual scores of finished tobacco of different specifications of cigarettes and the requirements for stability control accuracy; if The quality of the i-th batch is stable if Then the i-th batch is considered to have quality deviation.

3. The evaluation method according to claim 1 or 2, characterized in that: The recognition model consists of the entropy-weighted mean and the recognition interval.

4. The evaluation method according to claim 1 or 2, characterized in that: The key chemical component indicators of the model are total alkaloids, reducing sugars, total sugars, total nitrogen, K, chlorine, pH, and starch.

5. The evaluation method according to claim 1 or 2, characterized in that: This includes the use of a tobacco near-infrared big data system, in which the moisture content of the sample is modulated in the range of 6-8% before near-infrared scanning.

6. The evaluation method according to claim 1 or 2, characterized in that: The method comprises the following steps before scanning: crushing the sample into 40-80 meshes, and the sample is shredded tobacco, tobacco stems, or tobacco dust.

7. The evaluation method according to claim 6, characterized in that: Including the pre-scanning steps: crush the sample into 60 mesh.

Images