Method for quickly and efficiently predicting shelf life of chicken

By measuring and correlation analysis of various indicators of chicken, the content of gingerol is predicted by using electronic nose and LightGBM models, the problem of difficulty in identifying chicken shelf life in the existing technology is solved, and the rapid and efficient prediction of chicken spoilage is achieved, which improves the Determination efficiency and accuracy.

CN118961495AInactive Publication Date: 2024-11-15JIANGSU INST OF POULTRY SCI +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411051377.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing technology lacks effective methods to quickly and efficiently identify and predict the shelf life of chicken, resulting in poor quality, spoilage, and rotten meat when the chicken is stored for too long, affecting taste and health.

Method used

By measuring and correlating the quality, flavor, smell, tasting, sensory evaluation, metabolic group, flavor group, etc. of chicken in different shelf life, the gingerol in chicken is used as a marking indicator for rapid and efficient prediction of shelf life. Electronic nose determination of volatile organic compounds and LightGBM regression model combined with BOA butterfly optimization algorithm to predict the gingerol content to determine the shelf life of the chicken.

Benefits of technology

It realizes rapid and efficient prediction of the time of chicken spoilage, improves the efficiency and accuracy of the measurement, and ensures the reliability and accuracy of the prediction results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118961495A_ABST
    Figure CN118961495A_ABST
Patent Text Reader

Abstract

The invention discloses a method for quickly and efficiently predicting the shelf life of chicken, which is characterized in that the content of gingerol in the chicken in the shelf life is measured, and when the content of gingerol is remarkably increased, the chicken is judged to be unedible, so that the aim of quickly and efficiently predicting the shelf life of the chicken is fulfilled. According to the method, the deterioration time of the chicken can be quickly and efficiently predicted, and the determination efficiency and accuracy are improved. According to the method, the concentration of the volatile organic compounds is measured through the electronic nose, and the LightGBM regression model is combined, so that the shelf life of the chicken can be quickly and efficiently predicted. Data of volatile organic compounds in a chicken sample can be rapidly obtained through the electronic nose technology, and the data serve as input characteristics of the model and are processed through the LightGBM regression model, so that the content of gingerol in chicken can be accurately predicted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to an identification method for ensuring the quality of chicken meat during shelf life, and belongs to the field of meat quality during shelf life. Background Art

[0002] Fresh chicken has a certain shelf life. If the chicken is left for too long, the meat quality will deteriorate, deteriorate, and become corrupt, which will affect the taste of the chicken and people's health. However, there is currently a lack of evaluation indicators, methods, and devices for identifying the edible shelf life of chicken. The present invention measures and performs correlation analysis on the meat quality, flavor, smell, taste, sensory evaluation, metabolome, flavor group, etc. of chicken at different shelf lives, and uses gingerol in chicken as a marker indicator for quickly and efficiently predicting the shelf life of chicken. Summary of the invention

[0003] In order to further solve the deficiencies of the above-mentioned existing technologies and promote the development of non-destructive testing technology for chicken, realize rapid and efficient prediction of the deterioration time of chicken, and improve the efficiency and accuracy of measurement, a method for rapidly and efficiently predicting the shelf life of chicken is proposed.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] The invention discloses an intelligent prediction method for quickly and efficiently predicting the shelf life of chicken. The specific steps are as follows: by predicting and measuring the content of gingerol in chicken during the shelf life, when the gingerol content increases significantly, it is determined that the chicken is no longer edible, thereby achieving the purpose of quickly and efficiently predicting the shelf life of chicken.

[0006] Furthermore, before determining the gingerol content in chicken during the shelf life, a conventional meat quality determination is performed, and the specific process is as follows:

[0007] Conventional meat quality determination includes water loss rate, pH, and meat color determination methods:

[0008] The water loss rate was measured using a strain-controlled unconfined pressure instrument: a meat sample of about 0.5 g was taken from each chicken breast, weighed W1 on a balance, placed between two layers of medical gauze, with 18 layers of filter paper on each side, pressurized to 35 kg, maintained for 5 min, and after the pressure was removed, the meat sample was weighed W2, and the water loss rate = (W1-W2) / W1×100%;

[0009] The pH is measured directly using an acidometer: select three points along the midline of the long axis of the pectoralis major muscle from head to tail, insert the probe of the pH meter directly into the surface of the pectoralis major muscle to a depth of 0.5 to 1 cm, record the pH, and calculate the average of the three measured values;

[0010] Meat color determination: Brightness (L), redness (a), and yellowness (b) values ​​were determined using a colorimeter: within the middle 1 / 3 of the surface of the chicken breast close to the ribs, try to avoid connective tissue, blood stasis, and visible fat. The calibrated colorimeter lens was placed vertically on the meat surface with the lens tightly against the meat surface. Three points were measured for each sample.

[0011] Furthermore, the electronic nose measurement was carried out after the conventional meat quality measurement: after cutting the sample into small pieces with scissors, 1g of the sample was weighed and placed in a clean 100mL beaker, sealed with a double layer of plastic wrap, and allowed to stand at room temperature for 1h before being tested on the machine; 3 parallel samples were prepared for each sample, and the injection needle was directly inserted into the sealed beaker containing the sample, and the electronic nose was used for measurement; measurement conditions: sampling time was 1 second / group; sensor self-cleaning time was 60 seconds; sensor zeroing time was 5 seconds; sample preparation time was 5 seconds; injection flow rate was 200ml / min; analysis sampling time was 60 seconds.

[0012] Further, the specific steps for predicting and determining the content of gingerol in chicken during the shelf life are as follows:

[0013] Step 1, sample preparation:

[0014] Sampling: Take a 50 to 200 g sample from the refrigerated chicken;

[0015] Cutting: Cut the chicken samples into small pieces of 1 cm × 1 cm to increase the surface area and make it easier for volatile organic compounds to be released;

[0016] Weighing: Use a precision electronic balance to weigh about 10 grams of chicken pieces to ensure that the sample size is consistent each time;

[0017] Sample loading: Place the weighed chicken pieces in a clean, dry beaker;

[0018] Sealing: Use double-layer plastic wrap or a sealing lid to seal the beaker to prevent volatile organic compounds from escaping before measurement;

[0019] Standing: Let the sealed beaker stand at room temperature for 30 minutes to ensure that the volatile organic compounds are fully released and evenly distributed in the gas phase around the sample;

[0020] Step 2: Electronic nose measurement:

[0021] Set the electronic nose parameters:

[0022] Sampling time (Ts): indicates the time the electronic nose collects volatile organic compound gases in each measurement;

[0023] Sensor self-cleaning time (Tc): indicates the time it takes for the sensor to remove residual odor through the self-cleaning process after each measurement;

[0024] Sensor zeroing time (Tz): indicates the time it takes for the sensor to zero before each measurement to ensure a stable baseline;

[0025] Sample preparation time (Tp): refers to the preparation time from opening the sealed state of the sample to starting the measurement;

[0026] Injection flow rate (Fs): indicates the flow rate of the gas sample passing through the sensor during the measurement process;

[0027] Analysis sampling time (Ta): represents the time used to analyze and record the sensor response signal;

[0028] Experimental steps:

[0029] Sample handling: Open the sealed beaker and immediately place the sample on the sampling port of the electronic nose;

[0030] Injection: Start the electronic nose and introduce the sample gas into the sensor module according to the set injection flow rate;

[0031] Determination: During the sampling time, the electronic nose sensor captures and records the response signal of VOCs;

[0032] Data recording: After sampling, record the sensor response data and save it in CSV or other processable data format;

[0033] Self-cleaning: After completing a measurement, the sensor self-cleaning program is started to remove residual odor and ensure the accuracy of the next measurement;

[0034] Step 3: Data preprocessing:

[0035] Outlier removal:

[0036] X outlier ={x∈X||x-μ|>3σ}

[0037] X clean =X\X outlier

[0038] Among them, μ is the mean, σ is the standard deviation, and X outlier is a set of outliers, X is the original data set, x is a single response value, X clean is the cleaned data set;

[0039] Data Standardization:

[0040]

[0041] X normalized ={x normalized ∣x∈X clean}

[0042] Among them, x normalized is the normalized response value, min(X) is the minimum value in the data set, max(X) is the maximum value in the data set, and X normalized is the normalized data set;

[0043] Step 4: Feature extraction:

[0044] Extraction of volatile organic compounds concentration:

[0045] From the preprocessed data set, extract the volatile organic compound (VOC) concentration value at each time point. Assume that the data set obtained by the electronic nose is X, where X i Indicates the VOC concentration value of the i-th sample, where i is the sample number:

[0046] x max,i =max(X i )

[0047] t response,i = argmax t X i (t)

[0048]

[0049] Among them, x max,i is the maximum response value, t response,i is the response time, x stable,i is the stable response value, T stable is the duration of the stable response period, t end is the end time of the stable period, t start is the start time of the stable period, argmax t X i (t) is the time at which X i (t) the time point at which the maximum value is reached;

[0050] Principal Component Analysis (PCA):

[0051] First calculate the covariance matrix ∑:

[0052]

[0053] Where N represents the total number of samples. is the feature mean vector, z i represents the standardized eigenvector of the i-th sample, and T represents the matrix transpose;

[0054] Next is the eigenvalue decomposition:

[0055] Σv=λv

[0056] Among them, λ represents the eigenvalue, and v represents the corresponding eigenvector;

[0057] Then comes dimensionality reduction:

[0058] Select the eigenvectors corresponding to the first k eigenvalues, project the original high-dimensional data into the low-dimensional space, and obtain the reduced-dimensional data set:

[0059] X PCA =XV k

[0060] Among them, V k is the matrix consisting of the first k eigenvectors, X PCA Represents the dataset after dimensionality reduction;

[0061] Step 5: Model selection and training:

[0062] First, the sample data set is divided into a training set, a validation set, and a test set according to the proportion of the preprocessed data set and the corresponding gingerol content labels;

[0063] This patent selects the LightGBM regression model, which has good performance and efficiency when processing large-scale data and high-dimensional features;

[0064] Loss function:

[0065]

[0066] Model iteration, by iteratively constructing m trees to approximate the objective function:

[0067]

[0068] Among them, L(y,f(X)) is the loss function, y i is the gingerol content of the i-th sample, y is the label vector, indicating the gingerol content of each sample, f(X) is the model prediction function, h m (X) is the mth tree, λ m is the corresponding learning rate, f(X i ) is the predicted value;

[0069] Gradient boosting, each tree is fitted based on the residual of the previous round:

[0070]

[0071] in, is the residual, f (m-1) (X) represents the output of the first m-1 trees, Indicates the loss change rate at the current position;

[0072] The goal of a new round of tree generation is to find a tree that minimizes the sum of squares of the residuals:

[0073]

[0074] in, is the residual of the i-th sample in the m-th round, argmin h It refers to finding the parameter h, h(X i ) is the input feature X i The predicted value after the current tree h;

[0075] Model Update:

[0076] f (m) (X) = f (m-1) (X)+λ m h m (X)

[0077] f (m) (X) is the model prediction value of the mth round, λ m is the learning rate, which controls the contribution of each tree to the final prediction;

[0078] Step 6: Model optimization:

[0079] The BOA butterfly optimization algorithm is introduced to optimize the parameters of the LightGBM model. First, the population M is initialized, and each individual represents a set of LightGBM model parameters:

[0080] θ j =(num l eaves j ,learning r ate j ,n e Stimulators j )

[0081] (j=1,2,…,M)

[0082] θ j Represents a set of parameter combinations, num l eaves j is the parameter for the number of leaf nodes in the jth solution, learning r ate j is the learning rate parameter in the jth solution, n e Stimulators j is the number parameter of trees in the jth solution;

[0083] Next is the fitness evaluation, and the fitness function uses the mean square error (MSE):

[0084]

[0085] y val,i is the true gingerol content of the i-th validation set sample, is the predicted value of the i-th validation set sample under the j-th solution, N val is the total number of samples in the validation set, MSE j is the mean square error of the jth solution;

[0086] In addition, butterflies will constantly update their positions according to their fitness, tending to butterflies with better fitness, and butterflies will exchange information with each other to improve global search capabilities.

[0087]

[0088] in, is the position of the jth solution in the t+1th generation, is the position of the jth solution in the tth generation, α is the step size coefficient, which controls the update amplitude of the solution, S i is the location of the solution with better fitness;

[0089] Next, determine the optimal solution:

[0090]

[0091] During the iteration process, the solution with the best fitness is retained as the final LightGBM model parameter, θ * is the optimal solution, It means to find the variable value that minimizes the objective function (here MSE);

[0092] Finally, the new electronic nose measurement data X new Input into the optimized LightGBM model to obtain the predicted gingerol content To determine the shelf life of chicken, when the gingerol content increases significantly, the chicken is judged to be inedible, thereby achieving the purpose of quickly and efficiently predicting the shelf life of chicken.

[0093] Compared with the prior art, the present invention has the following beneficial effects:

[0094] 1. The present invention can quickly and efficiently predict the deterioration time of chicken, thereby improving the efficiency and accuracy of the measurement.

[0095] 2. The present invention measures the concentration of volatile organic compounds through an electronic nose, and combined with the LightGBM regression model, the present invention can quickly and efficiently predict the shelf life of chicken. Electronic nose technology can quickly obtain data on volatile organic compounds in chicken samples. These data are used as input features of the model. After being processed by the LightGBM regression model, the content of gingerol in chicken can be accurately predicted. In order to further improve the prediction accuracy of the model, the present invention introduces the BOA butterfly optimization algorithm to optimize the parameters of the LightGBM model. The BOA algorithm simulates the behavior of butterflies foraging and information exchange, has the ability of rapid convergence and global search, and significantly improves the prediction accuracy and stability of the model, thereby ensuring the reliability of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 This is a graph showing the difference in the changes of capsaicin in these chicken samples from day 0 to day 10;

[0097] Figure 2 This is the PCA analysis diagram of chicken samples at different times;

[0098] Figure 3 This is a radar chart of the main taste indicators of chicken;

[0099] Figure 4 The present invention is a flow chart of an intelligent prediction method for quickly and efficiently predicting the shelf life of chicken. DETAILED DESCRIPTION

[0100] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0101] The present invention measures and performs correlation analysis on the meat quality, flavor, smell, taste, sensory evaluation, metabolome, flavor group, etc. of chicken within different shelf lives, and uses gingerol in chicken as a marker for quickly and efficiently predicting the shelf life of chicken.

[0102] Embodiment 1:

[0103] 18 terminal paternal hens with a body weight variation coefficient within 10% were selected. After scalding and slaughter, the breast muscles on both sides were quickly collected, stored in a clean packaging bag, quickly cooled in an ice water bath, and stored at 4°C for 7 days (a special refrigerator was used, strictly stored at 4°C, and the refrigerator should not be opened or closed at will during storage). Samples were collected at 0 / 1 / 2 / 4 / 5 / 6 / 8 / 10 days of storage, and 10 samples were collected at each time point for sensory evaluation, conventional meat quality, volatile basic nitrogen, volatile flavor substances, metabolomics, electronic nose, and electronic tongue determination.

[0104] 1. Conventional meat quality determination: The determination methods of water loss rate, pH and meat color refer to "NYT 2793-2015 Objective Evaluation Method of Edible Quality of Meat"; water loss rate uses a strain-controlled unconfined pressure instrument: take a meat sample of about 0.5g from each chicken breast, weigh it on a balance (W1), place it between two layers of medical gauze, pad 18 layers of filter paper on the top and bottom, pressurize it with 35kg, keep it for 5 minutes, remove the pressure, weigh the meat sample (W2), water loss rate = (W1-W2) / W1×100%. pH is directly measured using an acidometer.

[0105] Determination method: Take 3 points along the midline of the long axis of the pectoralis major muscle from head to tail, insert the probe of the pH meter directly into the surface of the pectoralis major muscle at a depth of 0.5 to 1 cm, record the pH, and calculate the average of the 3 measured values. The brightness (L) value, redness (a) value, and yellowness (b) value are measured using a colorimeter. Determination method: In the middle 1 / 3 area of ​​the surface of the chicken breast near the ribs, try to avoid connective tissue, blood stasis and visible fat, and place the corrected colorimeter lens vertically on the meat surface, with the lens tightly against the meat surface (no light leakage). Measure 3 points for each sample.

[0106] 2. Determination of volatile basic nitrogen: Determination shall be carried out in accordance with GB 5009.228-2016 National Food Safety Standard “Determination of Volatile Basic Nitrogen in Foods”.

[0107] 3. Determination of flavor substances: The volatile organic compound content in the pectoral muscle was determined using a FlavourSpec 1H1-00053 headspace gas chromatograph-ion mobility spectrometer (purchased from GAS, Germany).

[0108] 4. Electronic nose measurement: The electronic nose uses the direct headspace aspiration method: cut the sample into small pieces with scissors, weigh 1g of the sample and place it in a clean 100mL beaker, seal it with double-layer plastic wrap, let it stand at room temperature for 1 hour, and then test it on the machine. Prepare 3 parallels for each sample. Directly insert the injection needle into the sealed beaker containing the sample, and the electronic nose performs the measurement. Measurement conditions: sampling time is 1 second / group; sensor self-cleaning time is 60 seconds; sensor zeroing time is 5 seconds; sample preparation time is 5 seconds; injection flow rate is 200ml / min; analysis sampling time is 60 seconds.

[0109] 5. Metabolome analysis: 500 mg of pectoral muscle was collected every day and stored in cryotubes at -80°C. The samples were sent to Panomik for metabolome analysis.

[0110] Results and analysis:

[0111] 1. If Figure 1 , simply analyzing the trend of gingerol (including gingerol) over time, it was found that the chicken deteriorated after 8 days of shelf life through the determination of chicken meat quality, volatile basic nitrogen, flavor, smell, taste, and sensory evaluation. Gingerol increased rapidly and significantly on the 8th day.

[0112] 2. Results of conventional meat quality test: The water loss rate on the shelf life of 0 days was significantly lower than that at other times (P≤0.05), and the pH of chicken on the 7th day of storage was significantly lower than that on the 0th and 2nd days. In general, the effect of chicken stored within 10 days on its conventional meat quality was not significant, as shown in Table 1:

[0113] Table 1

[0114] Water loss rate PH l a b 0 37.47±3.79b 5.91±0.08a 53.63±1.16 0.72±0.29 8.69±1 1 43.17±2.34ab 5.86±0.05ab 53.81±1.79 0.83±0.64 7.9±1.38 2 47.54±1.31a 5.91±0.05a 54.87±1.07 0.47±0.44 8.31±0.99 3 44.26±1.87a 5.82±0.06ab 53.93±0.99 1.35±0.25 8.72±1.18 4 43.79±1.09a 5.84±0.05ab 54.59±0.8 1.56±1.01 8.82±0.93 5 44.27±0.9a 5.85±0.04ab 55.96±1.59 0.35±0.54 7.9±1.41 6 43.61±1.9a 5.84±0.05ab 55.04±0.97 0.75±0.45 9.73±0.97 7 44.3±1.48a 5.71±0.04b 56.63±1.94 0.77±0.31 9.6±0.73 8 43.62±1.08a 5.82±0.04ab 55.69±0.88 0.53±0.18 8.5±0.63 10 43.14±1.73ab 5.86±0.06ab 54.98±1.36 0.3±0.37 10.42±1.05

[0115] 3. Volatile basic nitrogen test results: Generally speaking, the volatile basic nitrogen content of chicken increases with time during the shelf life. According to GB 2707-2016 "National Food Safety Standard for Fresh (Frozen) Livestock and Poultry Products", the volatile basic nitrogen values ​​of fresh chicken measured from 0 to 8 days are all ≤15mg / 100g, which is fresh meat. The figure shows that the TVB-N value has an inflection point at 8 days of storage. On the 10th day, the TVB-N value is (17.25±1.25) mg / 100g, which exceeds the range of fresh meat (<15mg / 100g), indicating that the chicken has deteriorated at this time.

[0116] 4. Results of volatile flavor substance determination: A total of 67 volatile substances were identified after chicken was stored for 0, 1, 2, 4, 5, 6, 8, and 10 days. According to the NIST mass spectrum library, there are 52 compounds that were identified, including 11 aldehydes, 10 alcohols, 15 ketones, 7 esters, and 9 other compounds, some of which include monomers and dimers. There are 15 unknown compounds including 1, 4, 6, 7, 9, 15, 25, 29, 31, 58, 138, 152, 184, 190, and 268.

[0117] Aldehydes are mainly derived from the oxidation of unsaturated fatty acids. They are highly volatile and contribute to the flavor of chicken, giving it an aroma. The identified aldehydes include butyraldehyde, valeraldehyde, hexanal, heptanal, benzaldehyde, octanal, nonanal and propanal. Among them, the concentration of n-butyraldehyde was the highest on the 6th day of storage, the concentrations of hexanal and valeraldehyde decreased with the extension of storage time, the concentration of benzaldehyde was high on the 1st, 2nd, 4th and 5th days of storage, and the lowest concentration on the 10th day as time extended, the concentration of octanal was high on the 2nd, 5th and 8th days of storage, the concentration of nonanal was the highest on the 8th day of storage, the concentration of propionaldehyde was high on the 6th and 8th days of storage, and the concentrations of 2-hexen-1-ol, 2-ethyl-1-hexanol and 1,8-cineole were high on the 10th day of storage. Relevant studies have shown that aldehydes are divided into unsaturated aldehydes and saturated aldehydes, among which unsaturated aldehydes, such as enal, are the main aroma components in chicken soup. Oleic acid produces nonanal and octanal after oxidation, which have oily aroma; linoleic acid and arachidonic acid produce hexanal and heptanal after oxidation. Hexanal has a grassy taste, heptanal has a nutty and sweet taste; valeraldehyde has a fruity aroma.

[0118] Alcohols also come from the oxidation of unsaturated fatty acids. Their odor threshold is higher than that of aldehydes, but they have unique and pleasant flavors such as floral and fruity aromas. Unsaturated alcohols have a lower threshold and have a great impact on the flavor of chicken. The identified alcohols include 1-propanol, n-butanol, 1-pentanol, 1-octen-3-ol (mushroom alcohol), n-hexanol, phenylethyl alcohol, 2-hexen-1-ol, 2-ethyl-1-hexanol, 1,8-cineole and n-octanol. As the storage period prolonged, the concentration of 1-propanol decreased, the concentration of n-butanol was higher on the 4th and 5th days, the concentration of 1-pentanol increased after the 4th day, the concentration of 1-octen-3-ol (mushroom alcohol) increased after the 4th day, and the concentration was highest on the 6th and 8th days. The concentrations of n-hexanol, 1,8-cineole and n-octanol were highest on the 10th day. n-Hexanol is produced by the oxidation of palmitic acid and oleic acid, and 1-octen-3-ol is mainly produced by the oxidation of linoleic acid. It has a typical mushroom flavor and fragrance, which can enhance the fat flavor of chicken soup.

[0119] Ketones are an important carbonyl compound in chicken soup. They are derived from the oxidation of unsaturated fatty acids and / or the degradation of amino acids. They generally have floral, fruity, and creamy aromas, but their thresholds are high, and their types are few and low, so they contribute little to the flavor of chicken soup. The ketones detected include 2,3-butanedione, 2-butanone, 2,3-pentanedione, 2-pentanone, acetophenone, furanone, 2-heptanone, 2-hexanone, 2-octanone, cyclohexanone, acetone, and 2-nonanone.

[0120] The degree of oxidation of chicken increased with the extension of storage time, and the concentrations of 2-pentanone, 2-heptanone, 2-hexanone and 2-nonanone increased on the 10th day of storage.

[0121] Esters are produced by the esterification reaction of alcohol and fatty acids, and have a fruity aroma, which has an important influence on the flavor of chicken soup. The ester substances detected include ethyl acetate, isopropyl acetate, ethyl propionate, ethyl acrylate, 3-methylbutyl acetate, (Z)-3-hexenyl pentanoate and isoamyl acetate. The esters produced by short-chain fatty acids have a typical fruity aroma, and the esters produced by long-chain fatty acids have a greasy smell. The concentrations of 3-methylbutyl acetate, (Z)-3-hexenyl pentanoate and isoamyl acetate are the highest on the 10th day of storage.

[0122] From the above, we can see that the volatile flavor substances change greatly after 10 days of storage.

[0123] 5. Electronic tongue test results: PCA was performed on chicken samples from 0 to 10 days based on all taste indicators. The results are shown in Figure 2 In order to better display similar samples, the same days are represented by the same color.

[0124] Connect the samples at the same time, from Figure 2 As can be seen, 0d to 6d (samples 1 to 21) overlap in the figure, and 8d and 10d can be completely distinguished from 0d to 6d. The difference is mainly reflected in the second principal component. Richness contributes the most to the second principal component. In addition, bitter aftertaste, astringent aftertaste, sweetness, etc. also have a greater function on the second principal component. It can be seen that the difference between 8d and 10d and 0d to 6d may be mainly reflected in these indicators.

[0125] like Figure 3 As shown, it can be seen that the richness of 8d and 10d chicken is concentrated together, with the lowest richness; the bitterness of 8d and 10d chicken samples is slightly greater, and the chicken samples from 0d to 6d are concentrated in the figure; the bitter aftertaste values ​​of 8d and 10d are significantly higher than those from 0d to 6d. The bitter aftertaste values ​​of 0d to 6d are all lower than the tasteless point, and the bitter aftertaste values ​​of 8d and 10d are slightly higher than the tasteless point; the sweetness of 8d and 10d samples is the lowest, and the sweetness of 0d to 6d samples is between 3.5 and 4, which is slightly higher than that of 8d and 10d.

[0126] 6. Electronic nose measurement results: The electronic nose measurement shows that the odor of the samples at 10 days is the most different from that at other times. A careful comparison of the samples from 0 days to 8 days shows that although the samples are mixed and distributed together in the figure, a certain distribution trend can be basically seen. The difference in odor between 0 days and 8 days is small and slow, while the difference between 10 days and other samples is huge, and the response values ​​of the main responding sensors are also much higher than those of other samples;

[0127] Further, if Figure 4 The specific calculation steps for predicting the content of gingerol in chicken during the shelf life are as follows:

[0128] Sampling: Take a 50 to 200 g sample from the refrigerated chicken;

[0129] Cutting: Cut the chicken samples into small pieces of 1 cm × 1 cm to increase the surface area and make it easier for volatile organic compounds to be released;

[0130] Weighing: Use a precision electronic balance to weigh about 10 grams of chicken pieces to ensure that the sample size is consistent each time;

[0131] Sample loading: Place the weighed chicken pieces in a clean, dry beaker;

[0132] Sealing: Use double-layer plastic wrap or a sealing lid to seal the beaker to prevent volatile organic compounds from escaping before measurement;

[0133] Standing: Let the sealed beaker stand at room temperature for 30 minutes to ensure that the volatile organic compounds are fully released and evenly distributed in the gas phase around the sample;

[0134] Set the parameters of the electronic nose:

[0135] Sampling time: 60 seconds;

[0136] Sensor self-cleaning time: 30 seconds;

[0137] Sensor zeroing time: 10 seconds;

[0138] Sample preparation time: 5 minutes;

[0139] Injection flow rate: 200 ml / min;

[0140] Analysis sampling time: 3 minutes

[0141] The preprocessed data set is used for training at 70%, validation at 15%, and testing at 15%. The initialization parameters of the LightGBM regression model are set as follows:

[0142] num_leaves: 31;

[0143] learning_rate: 0.05;

[0144] n_estimators: 100;

[0145] Use the BOA butterfly optimization algorithm to optimize the parameters of the LightGBM model, and set them as follows:

[0146] Population size: 20;

[0147] Step size coefficient (α): 0.1;

[0148] Initial parameter range: num_leaves: 20 to 50;

[0149] learning_rate: 0.01 to 0.1;

[0150] n_estimators: 50 to 200;

[0151] The following are some of the values ​​and actual physical and chemical values ​​predicted by the intelligent prediction algorithm after using the electronic nose to measure the data and dividing the data set into 70% for training, 15% for verification, and 15% for testing:

[0152] 0d: gingerol content (true value): 10 mg / kg, predicted value: 9.8 mg / kg;

[0153] 2d: gingerol content (true value): 15 mg / kg, predicted value: 14.7 mg / kg;

[0154] 4d: gingerol content (true value): 20 mg / kg, predicted value: 19.5 mg / kg;

[0155] 6d: gingerol content (actual value): 25 mg / kg, predicted value: 24.8 mg / kg;

[0156] 8d: gingerol content (actual value): 30 mg / kg, predicted value: 29.7 mg / kg;

[0157] 10d: gingerol content (actual value): 50 mg / kg, predicted value: 49.2 mg / kg;

[0158] It can be seen that the intelligent prediction algorithm of the present invention is relatively accurate in the regression prediction of the true value, and the final evaluation index RMSEP of the entire data set can reach 0.28 mg / kg.

[0159] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for quickly and efficiently predicting the shelf life of chicken, characterized by: By predicting and measuring the gingerol content in chicken during its shelf life, when the gingerol content increases significantly, the chicken is judged to be inedible, thereby achieving the purpose of quickly and efficiently predicting the shelf life of chicken.

2. The method according to claim 1, characterized in that: Before determining the gingerol content in chicken during the shelf life, a conventional meat quality test is performed. The specific process is as follows: Conventional meat quality determination includes water loss rate, pH, and meat color determination methods: The water loss rate was measured using a strain-controlled unconfined pressure instrument: a meat sample of about 0.5 g was taken from each chicken breast, weighed W1 on a balance, placed between two layers of medical gauze, with 18 layers of filter paper on each side, pressurized to 35 kg, maintained for 5 min, and after the pressure was removed, the meat sample was weighed W2, and the water loss rate = (W1-W2) / W1×100%; The pH is measured directly using an acidometer: select three points along the midline of the long axis of the pectoralis major muscle from head to tail, insert the probe of the pH meter directly into the surface of the pectoralis major muscle to a depth of 0.5 to 1 cm, record the pH, and calculate the average of the three measured values; Meat color determination: Brightness (L), redness (a), and yellowness (b) values ​​were determined using a colorimeter: within the middle 1 / 3 of the surface of the chicken breast close to the ribs, try to avoid connective tissue, blood stasis, and visible fat. The calibrated colorimeter lens was placed vertically on the meat surface with the lens tightly against the meat surface. Three points were measured for each sample.

3. The method according to claim 2, characterized in that: After the conventional meat quality test, the electronic nose test was performed: the sample was cut into small pieces with scissors, 1g of the sample was weighed and placed in a clean 100mL beaker, sealed with double-layer plastic wrap, and left to stand at room temperature for 1h before testing on the machine; 3 parallel samples were prepared for each sample, and the injection needle was directly inserted into the sealed beaker containing the sample, and the electronic nose was used for the test; Measurement conditions: sampling time is 1 second / group; sensor self-cleaning time is 60 seconds; sensor zeroing time is 5 seconds; sample preparation time is 5 seconds; injection flow rate is 200 ml / min; analysis sampling time is 60 seconds.

4. The method according to claim 1, characterized in that: The specific steps for predicting and determining the content of gingerol in chicken during the shelf life are as follows: Step 1, sample preparation: Sampling: Take a 50 to 200 g sample from the refrigerated chicken; Cutting: Cut the chicken samples into small pieces of 1 cm × 1 cm to increase the surface area and make it easier for volatile organic compounds to be released; Weighing: Use a precision electronic balance to weigh about 10 grams of chicken pieces to ensure that the sample size is consistent each time; Sample loading: Place the weighed chicken pieces in a clean, dry beaker; Sealing: Use double-layer plastic wrap or a sealing lid to seal the beaker to prevent volatile organic compounds from escaping before measurement; Standing: Let the sealed beaker stand at room temperature for 30 minutes to ensure that the volatile organic compounds are fully released and evenly distributed in the gas phase around the sample; Step 2: Electronic nose measurement: Set the electronic nose parameters: Sampling time Ts: represents the time that the electronic nose collects volatile organic compound gases in each measurement; Sensor self-cleaning time Tc: indicates the time it takes for the sensor to remove residual odor through the self-cleaning process after each measurement; Sensor zeroing time Tz: indicates the time for the sensor to return to zero before each measurement to ensure a stable baseline; Sample preparation time Tp: refers to the preparation time from opening the sealed state of the sample to starting the measurement; Injection flow rate Fs: indicates the flow rate of the gas sample passing through the sensor during the measurement process; Analysis sampling time Ta: represents the time used to analyze and record the sensor response signal; Experimental steps: Sample handling: Open the sealed beaker and immediately place the sample on the sampling port of the electronic nose; Injection: Start the electronic nose and introduce the sample gas into the sensor module according to the set injection flow rate; Determination: During the sampling time, the electronic nose sensor captures and records the response signal of VOCs; Data recording: After sampling, the sensor response data is recorded and saved in CSV data format; Self-cleaning: After completing a measurement, the sensor self-cleaning program is started to remove residual odor and ensure the accuracy of the next measurement; Step 3: Data preprocessing: Outlier removal: X outlier ={x∈X∣|x-μ|>3σ} X clean =X\X outlier Among them, μ is the mean, σ is the standard deviation, and X outlier is a set of outliers, X is the original data set, x is a single response value, X clean is the cleaned data set; Data Standardization: X normalized ={x normalized ∣x∈X clean } Among them, x normalized is the normalized response value, min(X) is the minimum value in the data set, max(X) is the maximum value in the data set, and X normalized is the normalized data set; Step 4: Extract volatile organic compound concentration: From the preprocessed data set, extract the volatile organic compound (VOC) concentration value at each time point. Assume that the data set obtained by the electronic nose is X, where X i Indicates the VOC concentration value of the i-th sample, where i is the sample number: x max,i =max(X i ) t response,i =argmax t X i (t) Among them, x max,i is the maximum response value, t response,i is the response time, x stable,i is the stable response value, T stable is the duration of the stable response period, t end is the end time of the stable period, t start is the start time of the stable period, argmax t X i (t) is the time at which X i (t) the time point at which the maximum value is reached; Principal Component Analysis PCA: First calculate the covariance matrix δ: Where N represents the total number of samples. is the feature mean vector, z i represents the standardized eigenvector of the i-th sample, and T represents the matrix transpose; Next is the eigenvalue decomposition: ∑v=λv Among them, λ represents the eigenvalue, and v represents the corresponding eigenvector; Then comes dimensionality reduction: Select the eigenvectors corresponding to the first k eigenvalues, project the original high-dimensional data into the low-dimensional space, and obtain the reduced-dimensional data set: X PCA =XV k Among them, V k is the matrix consisting of the first k eigenvectors, X PCA Represents the dataset after dimensionality reduction; Step 5, first divide the sample data set, and divide the preprocessed data set and the corresponding gingerol content label into a training set, a validation set, and a test set according to the proportion; the present invention selects the LightGBM regression model, which has good performance and high efficiency in processing large-scale data and high-dimensional features; Loss function: Model iteration, by iteratively constructing m trees to approximate the objective function: Among them, L(y,f(X)) is the loss function, y i is the gingerol content of the i-th sample, y is the label vector, indicating the gingerol content of each sample, f(X) is the model prediction function, h m (X) is the mth tree, λ m is the corresponding learning rate, f(X i ) is the predicted value; Gradient boosting, each tree is fitted based on the residual of the previous round: in, is the residual, f (m-1) (X) represents the output of the first m-1 trees, Represents the rate of change of loss at the current position; the goal of a new round of tree generation is to find a tree that minimizes the sum of squares of the residual: in, is the residual of the i-th sample in the m-th round, argmin h It refers to finding the parameter h, h(X i ) is the input feature X i The predicted value after the current tree h; Model Update: f (m) (X)=f (m-1) (X)+λ m h m (X) f (m) (X) is the model prediction value of the mth round, λ m is the learning rate, which controls the contribution of each tree to the final prediction; Step 6: Introduce the BOA butterfly optimization algorithm to optimize the parameters of the LightGBM model. First, initialize the population M. Each individual represents a set of LightGBM model parameters: θ j =(num l eaves j ,learning r ate j ,n e stimators j ) (j=1,2,…,M) θ j Represents a set of parameter combinations, num l eaves j is the parameter for the number of leaf nodes in the jth solution, learning r ate j is the learning rate parameter in the jth solution, n e Stimulators j is the number parameter of trees in the jth solution; Next is the fitness evaluation, and the fitness function uses the mean square error MSE: y val,i is the true gingerol content of the i-th validation set sample, is the predicted value of the i-th validation set sample under the j-th solution, N val is the total number of samples in the validation set, MSE j is the mean square error of the jth solution; In addition, butterflies will constantly update their positions according to their fitness, tending to butterflies with better fitness, and butterflies will exchange information with each other to improve global search capabilities. in, is the position of the jth solution in the t+1th generation, is the position of the jth solution in the tth generation, α is the step size coefficient, which controls the update amplitude of the solution, S i is the location of the solution with better fitness; Next, determine the optimal solution: During the iteration process, the solution with the best fitness is retained as the final LightGBM model parameter, θ * is the optimal solution, It means finding the variable value that minimizes the objective function; Finally, the new electronic nose measurement data X new Input into the optimized LightGBM model, we get The predicted gingerol content To determine the shelf life of chicken, when the gingerol content increases significantly, Determine that the chicken is no longer edible, thereby achieving the goal of quickly and efficiently predicting the shelf life of chicken.