Cigarette quality control chart pattern recognition method based on interpretable graph convolutional neural network

Abstract

The present invention relates to the technical field of cigarette production monitoring, and more particularly to a method for pattern recognition of cigarette quality control charts based on an interpretable graph convolutional neural network. The method comprises: obtaining multiple cigarette control chart training samples under different modes; extracting training features of each cigarette control chart training sample and generating a feature test set; constructing multiple classification network models, inputting the feature test set corresponding to each cigarette control chart training sample into the multiple classification network models, and obtaining multiple classification prediction result data and ROC curve graphs; screening the optimal classification network model based on the ROC curve graph and inputting it into a SHAP interpreter; calculating the marginal contribution of each feature in the feature test set to the classification prediction result data, and generating a high-value feature subset based on the Shapley value; using the high-value feature subset to train and obtain a target graph convolutional neural network; obtaining actual collected cigarette control chart samples, and inputting them into the target graph convolutional neural network to complete pattern recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of cigarette production monitoring, and in particular to a control chart pattern recognition method. Background Art

[0002] With the rapid development of intelligent manufacturing and information technology, control chart anomaly pattern recognition has become an effective method for intelligent quality monitoring of production processes. In the cigarette industry, real-time collection and analysis of cigarette quality data during the production process are commonly used to create control charts to identify anomaly patterns and monitor the operational status of the production process. This, combined with a fault database, facilitates tracing, accelerates diagnosis, and enables rapid recovery of the production process.

[0003] The quality of cigarettes is one of the important indicators to measure whether the final cigarette product is qualified or not. It can reflect the influence of various factors such as the process parameters of the cigarette rolling equipment and the processing technology during the processing of cigarettes. Therefore, identifying abnormal patterns in the cigarette quality control chart plays an important role in production process monitoring.

[0004] In the process of identifying abnormal control chart patterns, machine learning methods are needed to classify different control chart abnormal patterns based on features. Existing methods often focus on feature extraction to improve recognition accuracy. However, the contribution of extracted features to the classification results is often difficult to accurately quantify, which restricts the scientific and effective quality assessment of cigarettes. Summary of the Invention

[0005] The purpose of the present invention is to provide a cigarette quality control chart pattern recognition method based on an interpretable graph convolutional neural network. This method can interpret the shape characteristics or statistical characteristics of the control chart, thereby specifying different features for pattern classification, and can accurately quantify the contribution of various different feature quantities to the classification model results, thereby improving the scientificity and effectiveness of cigarette quality assessment.

[0006] In order to achieve the above object, the technical solutions adopted by the present invention are as follows:

[0007] A cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network is characterized in that the method comprises:

[0008] S01. Obtaining a plurality of cigarette control chart training samples under different modes;

[0009] S02. Extracting training features of each cigarette control chart training sample; the training features include multiple shape features and multiple statistical features;

[0010] S03, generating a feature test set based on the training features; constructing multiple classification network models, inputting the feature test set corresponding to each cigarette control chart training sample into the multiple classification network models, obtaining multiple classification prediction result data, and correspondingly obtaining ROC curve graphs;

[0011] S04. Screening and obtaining an optimal classification network model according to the ROC curve graph; inputting the optimal classification network model, the feature test set, and the classification prediction result data output by the optimal classification network model into a SHAP interpreter;

[0012] S05, the SHAP interpreter calculates the marginal contribution of each feature in the feature test set to the classification prediction result data, generates multiple Shapley values; and generates a high-value feature subset based on the Shapley values;

[0013] S06. Inputting the high-value feature subset into the constructed graph convolutional neural network, training the model and obtaining a target graph convolutional neural network;

[0014] S07. Obtain samples of the actual collected cigarette control graph and input them into the target graph convolutional neural network to perform pattern recognition and obtain cigarette quality fluctuation results.

[0015] Therefore, this method first requires obtaining multiple training samples of cigarette control charts under different patterns. However, since abnormal pattern data is difficult to obtain in batches during actual production, and destructive experiments are often required to expand the sample size, resulting in high data acquisition costs and poor practicality, the present invention proposes to use a Monte Carlo simulation algorithm (MC) to design different parameters and generate different types of mean trend charts as the original training samples for the model. Secondly, the quality characteristics of the cigarettes are analyzed to determine which continuous variables should be controlled using variable control charts, and the conventional and effective control chart, the mean control chart, is selected.

[0016] According to different abnormal factors, the mean control chart can be divided into several different modes, such as normal mode (NOR), cyclic mode (CYC), downward trend mode (DT), upward trend mode (UT), downward step mode (DS) and upward step mode (US).

[0017] Then, training features are extracted for each cigarette control chart training sample; the training features include multiple shape features and multiple statistical features. The present invention uses nine conventional and highly efficient shape features, namely: the slope of the least squares line, the average number of crossings (nC1), the number of least squares line crossings (nC2), the cyclic member (cMember), the average slope of the line segment (as), the slope difference (sD) between the least squares line and the line segment representing the pattern, the area between the pattern and the mean line (aPM), the area between the pattern and its least squares line (apS), and the area between the least squares line and the line segment (ass); as well as six statistical features, namely, the mean, the cumulative sum statistic (cusum), the skewness, the kurtosis, the mean square value, and the standard deviation.

[0018] Then, a feature test set is generated based on the training features; each feature test set corresponds to a cigarette control chart training sample, which includes 15 features extracted from the above cigarette control chart training sample.

[0019] Next, multiple classification network models are constructed, such as classic machine learning classification network models such as decision tree, random forest, support vector machine, GBDT, XGBoost and logistic regression, for preliminary classification; specifically, the feature test sets corresponding to each cigarette control chart training sample are input into multiple classification network models respectively, and multiple classification prediction result data are obtained, and corresponding ROC curve graphs are obtained.

[0020] The ROC curve contains multiple curves, each curve represents a different classification network model. The larger the area under the curve, the better the classification effect of the model. Therefore, the optimal classification network model can be screened according to the ROC curve.

[0021] Afterwards, the above-screened optimal classification network model, feature test set and classification prediction result data output by the optimal classification network model are input into the SHAP interpreter.

[0022] The SHAP interpreter calculates the marginal contribution of each feature in the feature test set to the classification prediction data, generating multiple Shapley values. Based on these Shapley values, a subset of high-value features is generated. SHAP, based on the Shapley value principle in game theory, calculates the marginal contribution of each feature to the classification prediction data of the optimal classification network model. The Shapley value quantifies the actual impact of a feature on the model output in addition to other features, while also accounting for interactions between features.

[0023] The high-value feature subset is then fed into the constructed graph convolutional neural network (GCNN) to train the model and obtain the target GCNN. GCNNs are a type of neural network used in the prior art to process graph-structured data. They can effectively capture the connections between nodes (i.e., samples). For control graph data, similarities between samples can be constructed as graph edges, and the classification result Y is ultimately output through an activation function.

[0024] Finally, the filtered features and trained model are used to monitor the dynamic data stream of cigarette quality. Specifically, samples of actual cigarette control charts are collected and fed into the target graph convolutional neural network for pattern recognition, identifying abnormal quality fluctuations in the cigarette production process.

[0025] In summary, in this method, with the help of the SHAP interpreter, the shape characteristics and statistical characteristics of the control chart were interpreted, the Shapley value was obtained, the contribution of different features to the classification model results was accurately quantified, and a high-value feature subset was generated. The model was trained with this high-value feature subset and the target graph convolutional neural network was obtained, which greatly improved the prediction accuracy of the target graph convolutional neural network for abnormal pattern recognition of cigarette quality control charts, and ultimately comprehensively improved the scientificity and effectiveness of cigarette quality assessment.

[0026] As a preferred embodiment of the present invention, in S01, the method for obtaining cigarette control chart training samples under multiple different modes is specifically: based on the control chart mode simulation formula, using the Monte Carlo method to simulate and obtain cigarette control chart training samples under multiple different modes.

[0027] Since abnormal pattern data is difficult to obtain in batches during actual production processes, and most of them require destructive experiments to expand the sample size of abnormal data, which leads to practical problems such as high data acquisition cost and poor practicality, the present invention intends to use the Monte Carlo simulation algorithm (MC) to design different parameters to generate different types of mean trend charts as the original training samples of the model.

[0028] As a preferred embodiment of the present invention, the modes of the cigarette control chart training samples include: normal mode (NOR), cycle mode (CYC), downward trend mode (DT), upward trend mode (UT), downward step mode (DS) and upward step mode (US).

[0029] Among them, normal mode (NOR): indicates that the machine is properly maintained, the material quality is stable, there are no major fluctuations, and it is in a stable state; cyclical mode (CYC): cyclical fluctuations of equipment, such as periodic mechanical wear or vibration of the rollers, gears or transmission devices of the rolling machine group, resulting in periodic fluctuations in cigarette quality; downward trend mode (DT) and upward trend mode (UT): changes in tobacco volume due to gradual wear of machine parts, or increased environmental humidity, resulting in a gradual decrease or increase in cigarette quality; downward step mode (DS) and upward step mode (US): indicate sudden changes in temperature or humidity of environmental factors, resulting in changes in tobacco moisture, sudden adjustments to equipment or changes in raw material batches, resulting in changes in cigarette indicators.

[0030] As a preferred embodiment of the present invention, the shape features include: the slope of the least squares line (slope), the average number of crossings (nc1), the number of least squares line crossings (nc2), the cyclic member (cmember), the average slope of the line segment (as), the slope difference (sd) between the least squares line and the line segment representing the pattern, the area between the pattern and the mean line (apm), the area between the pattern and its least squares line (aps), and the area between the least squares line and the line segment (ass).

[0031] Shape features provide information about the appearance and form of a graphic by describing its geometry and structure. They can help qualitatively describe and analyze the specific shape of a graphic. The nine shape features used in this invention are as follows:

[0032] The slope of the least squares line indicates the trend of the pattern. For normal and periodic patterns, the absolute value of the slope is close to zero, while for step and trend patterns, the absolute value of the slope is greater than zero.

[0033] Average number of crossings (nc1): indicates the number of intersections between the pattern and the mean line. The normal pattern has the most intersections, the step and trend patterns have the least, and the cyclical pattern is in between.

[0034] Number of least squares line crossings (nc2): indicates the number of times the pattern crosses the least squares line;

[0035] Cyclic member (cmember): Indicates the degree of similarity between the pattern and the periodic pattern. For periodic patterns, this value is positive, and for other patterns it is negative.

[0036] Average slope of the line segment (as): represents the average slope of each line segment in the pattern. The trend pattern has the highest average slope.

[0037] Slope difference (sd) between the least squares line and the segment representing the pattern: the difference between the slope of the least squares line and the average slope of the segment;

[0038] Area between the mode and the mean line (apm): The area of ​​this feature is lowest for the normal mode, so this feature is different between the normal mode and other modes;

[0039] Area between the pattern and the least squares line (aps): This feature has higher aps values ​​in periodic and step patterns than in normal and trend patterns, so aps can distinguish patterns such as periodic and step patterns;

[0040] The area between the least squares line and the line segment (ass): This feature is approximately 0 in the trend mode and may be higher for other modes, so this feature can distinguish the trend mode from other modes.

[0041] As a preferred embodiment of the present invention, the statistical features include: mean, cumulative sum statistic (cusum), skewness, kurtosis, mean square value (mean square value) and standard deviation (standard deviation).

[0042] Statistical features provide information such as data distribution, central tendency, and dispersion through mathematical calculations, which can help quantitatively describe and analyze the overall properties of the graph. The six statistical features used in this invention are as follows:

[0043] Mean: represents the location of the data center, used to determine whether the data fluctuates around a certain central value and show the central trend of the data;

[0044] Cumulative sum statistic (cusum): A statistic that detects changes by accumulating deviations between data points and target values. It is used to monitor process changes, especially small, persistent shifts. A positive cumulative sum statistic is used to detect upward mean shifts, and a negative cumulative sum statistic is used to detect downward mean shifts.

[0045] Skewness: reflects the symmetry of the data distribution. Positive skewness indicates that the data is right-skewed, and negative skewness indicates that the data is left-skewed. It is used to identify the process trend of the tilt, deviation or asymmetry of the data distribution.

[0046] Kurtosis is used to describe the sharpness of the data distribution. A high kurtosis indicates that the data distribution has a sharp peak, while a low kurtosis indicates a flat distribution. It is used to detect outliers or extreme values.

[0047] The mean square value is the average of the square values ​​of all data points. It is used to measure the energy or variability of a data set and can help detect specific trends in data. In particular, in control charts, it reflects volatility and instability. Larger values ​​indicate greater surface fluctuations.

[0048] The standard deviation indicates the degree of dispersion between data points and the mean, and is used to measure data fluctuation and dispersion. Compared with the mean square value, the standard deviation reflects more about how the data fluctuates around the mean.

[0049] As a preferred embodiment of the present invention, in S03, the multiple classification network models constructed are specifically: a decision tree model, a random forest model, a logistic regression model, a support vector machine model, an XGBoost model and a GBDT model.

[0050] As a preferred embodiment of the present invention, in S05, the specific method for generating a high-value feature subset based on the Shapley value is: calculating the average absolute value of each Shapley value and sorting them according to the cumulative proportion, and selecting features with the top 90% cumulative proportion of Shapley values ​​to form a high-value feature subset.

[0051] As a preferred embodiment of the present invention, in S06, the constructed graph convolutional neural network is trained using a back propagation algorithm to obtain the target graph convolutional neural network.

[0052] In summary, the present invention has the following beneficial effects:

[0053] In this method, the SHAP interpreter is used to interpret the shape and statistical characteristics of the control chart, obtain the Shapley value, accurately quantify the contribution of the feature quantity to the classification model results, and generate a high-value feature subset. This high-value feature subset is used to train the model and obtain the target graph convolutional neural network, which greatly improves the prediction accuracy of the target graph convolutional neural network in identifying abnormal patterns in cigarette quality control charts, ultimately comprehensively improving the scientificity and effectiveness of cigarette quality assessment.

[0054] The contribution differences of 9 shape features and 6 statistical features to different control chart patterns are studied, with a wide range of considerations, thereby improving the scientific nature of the features selected by the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A flow chart of the invented method;

[0056] Figure 2 This is a control diagram in the normal mode of this embodiment;

[0057] Figure 3 This is a control diagram in the periodic mode of this embodiment;

[0058] Figure 4 This is the control diagram in the upward step mode in this embodiment;

[0059] Figure 5 This is the control diagram in the downward step mode in this embodiment;

[0060] Figure 6 This is the control chart under the upward trend mode in this embodiment;

[0061] Figure 7 This is the control chart under the downward trend mode in this embodiment;

[0062] Figure 8 It is the ROC curve diagram in this embodiment;

[0063] Figure 9 Schematic diagram of a bee colony in this embodiment;

[0064] Figure 10 : is a topological structure diagram of the graph convolutional neural network in this embodiment;

[0065] Figure 11 : is the loss curve diagram in this embodiment;

[0066] Figure 12 is the confusion matrix diagram in this embodiment;

[0067] Figure 13 This is the cigarette quality mean control chart in this embodiment. DETAILED DESCRIPTION

[0068] The following is an explanation and description of the technical solutions of the embodiments of the present invention in conjunction with the drawings of the embodiments of the present invention. However, the following embodiments are only preferred embodiments of the present invention and are not exhaustive. Based on the embodiments in the implementation manner, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention.

[0069] Throughout this specification, the claims, and the accompanying drawings, the terms "first," "second," and so forth are used to distinguish between different items, not to describe a particular order. Furthermore, the term "comprises" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may include other steps or elements inherent to the process, method, product, or apparatus.

[0070] First, if Figure 1 As shown, multiple cigarette control chart training samples under different modes are obtained.

[0071] Specifically, a measurement value control chart, i.e., a mean control chart, is selected based on the characteristics of cigarette quality. Since it is difficult to obtain abnormal pattern data in batches during actual production, and destructive experiments are often required to expand the sample size of abnormal data, resulting in high data acquisition costs and poor practicality, the present invention proposes to use a Monte Carlo simulation algorithm (MC) to design different parameters to generate different types of mean trend charts, such as Figure 2-7 The graphs of the six modes are shown, and the simulation formulas for the different modes are as follows:

[0072] Where x(t) is the product's quality characteristic value at time t, u is the statistical mean under controlled conditions, n is the total number of time points, and σ is the process capability standard deviation. γ is the fluctuation caused by random factors, which follows a Gaussian white noise sequence and cannot be removed or is very costly to remove. d(t) is the fluctuation caused by abnormal factors at time t, the cause of which can be identified and eliminated. The simulation formulas for each mode are shown in Table 1:

[0073] Table 1 Simulation formula for each mode

[0074]

[0075]

[0076] Based on the above control chart pattern simulation formula, the embodiment of the present invention utilizes the Monte Carlo method to simulate data for six patterns. Furthermore, based on the frequency of collecting a cigarette sample every 6 seconds in the actual production line, a 5-minute time span (50 cigarettes in 300 seconds) is used to draw the control chart, i.e., a control chart sample is drawn with 50 quality characteristic observation points. The advantage of using a 5-minute time span for a control chart is that it can fully capture short-term fluctuations in the production process, improve the sensitivity of detecting anomalies, and facilitate early detection and correction of potential abnormal patterns. Therefore, during the Monte Carlo simulation, each control chart sample has 50 observation points, and 200 samples are simulated for each pattern, for a total of 1,200 samples. 80% of these samples are selected as the training set, and the remainder as the test set.

[0077] According to different abnormal factors, the mean control chart can be divided into several different modes, specifically including:

[0078] Normal mode (NOR): indicates that the machine is properly maintained, the material quality is stable, there are no major fluctuations, and it is in a stable state;

[0079] Cyclic mode (CYC): Cyclic fluctuations in equipment, such as periodic mechanical wear or vibration of the rollers, gears, or transmission devices of the cigarette making unit, leading to periodic fluctuations in cigarette quality;

[0080] Downward Trend Mode (DT) Upward Trend Mode (UT): This is due to the gradual wear of machine parts causing changes in tobacco volume, or increased ambient humidity leading to a gradual decrease or increase in cigarette mass;

[0081] Downward step mode (DS) and upward step mode (US): Indicates that sudden changes in temperature or humidity of environmental factors lead to changes in tobacco moisture, sudden adjustments to equipment or changes in raw material batches, which in turn lead to changes in cigarette indicators.

[0082] Then, the training features of each cigarette control chart training sample are extracted; the training features include 9 shape features and 6 statistical features.

[0083] Among them, shape features provide information about the appearance and form of a graphic by describing its geometric form and structure, which can help qualitatively describe and analyze the specific shape of the graphic. The nine shape features used in this invention are as follows:

[0084] The slope of the least squares line indicates the trend of the pattern. For normal and periodic patterns, the absolute value of the slope is close to zero, while for step and trend patterns, the absolute value of the slope is greater than zero.

[0085] Average number of crossings (nc1): indicates the number of intersections between the pattern and the mean line. The normal pattern has the most intersections, the step and trend patterns have the least, and the cyclical pattern is in between.

[0086] Number of least squares line crossings (nc2): indicates the number of times the pattern crosses the least squares line;

[0087] Cyclic member (cmember): Indicates the degree of similarity between the pattern and the periodic pattern. For periodic patterns, this value is positive, and for other patterns it is negative.

[0088] Average slope of the line segment (as): represents the average slope of each line segment in the pattern. The trend pattern has the highest average slope.

[0089] Slope difference (sd) between the least squares line and the segment representing the pattern: the difference between the slope of the least squares line and the average slope of the segment;

[0090] Area between the mode and the mean line (apm): The area of ​​this feature is lowest for the normal mode, so this feature is different between the normal mode and other modes;

[0091] Area between the pattern and the least squares line (aps): This feature has higher aps values ​​in periodic and step patterns than in normal and trend patterns, so aps can distinguish patterns such as periodic and step patterns;

[0092] The area between the least squares line and the line segment (ass): This feature is approximately 0 in the trend mode and may be higher for other modes, so this feature can distinguish the trend mode from other modes.

[0093] In addition, six statistical features are extracted from each cigarette control chart training sample, specifically:

[0094] Mean: represents the location of the data center, used to determine whether the data fluctuates around a certain central value and show the central trend of the data;

[0095] Cumulative sum statistic (cusum): A statistic that detects changes by accumulating deviations between data points and target values. It is used to monitor process changes, especially small, persistent shifts. A positive cumulative sum statistic is used to detect upward mean shifts, and a negative cumulative sum statistic is used to detect downward mean shifts.

[0096] Skewness: reflects the symmetry of the data distribution. Positive skewness indicates that the data is right-skewed, and negative skewness indicates that the data is left-skewed. It is used to identify the process trend of the tilt, deviation or asymmetry of the data distribution.

[0097] Kurtosis is used to describe the sharpness of the data distribution. A high kurtosis indicates that the data distribution has a sharp peak, while a low kurtosis indicates a flat distribution. It is used to detect outliers or extreme values.

[0098] The mean square value is the average of the square values ​​of all data points. It is used to measure the energy or variability of a data set and can help detect specific trends in data. In particular, in control charts, it reflects volatility and instability. Larger values ​​indicate greater surface fluctuations.

[0099] The standard deviation indicates the degree of dispersion between data points and the mean, and is used to measure data fluctuation and dispersion. Compared with the mean square value, the standard deviation reflects more about how the data fluctuates around the mean.

[0100] Then, a feature test set is generated based on the training features; each feature test set corresponds to a cigarette control chart training sample, which includes 15 features extracted from the above cigarette control chart training sample.

[0101] Next, multiple classification network models are constructed.

[0102] In order to avoid the bias caused by the prediction of a single model, six conventional machine learning models were selected for preliminary classification prediction.

[0103] Traditional machine learning models mainly focus on how to improve the prediction accuracy of the model, but ignore the interpretability of the prediction results. Therefore, the SHAP interpretable model can be used to quantitatively explain the contribution of shape features and statistical features to the control chart pattern recognition ability.

[0104] The classification network models in this embodiment are specifically: decision tree, random forest, support vector machine, GBDT, XGBoost and logistic regression machine learning classification network models.

[0105] Specifically, we first train samples of cigarette control charts under multiple different modes; divide the training set and test set into a ratio of 8:2 and input them into the designated models of decision tree, random forest, logistic regression, support vector machine, XGBoost and GBDT for preliminary classification. These six algorithms represent different learning methods and data processing capabilities and are suitable for comprehensive analysis of feature contribution. After training the models, we compare them and obtain a comparison chart of ROC curves of multiple models, as shown in the figure below. Figure 8 shown.

[0106] The ROC curve contains multiple curves, each curve represents a different classification network model. The larger the area under the curve, the better the classification effect of the model. Figure 8 It can be seen that by comparing the ROC curve graph, it can be concluded that the support vector machine is used as the optimal classification network model to input the subsequent calculation of the SHAP model.

[0107] Afterwards, the above-screened support vector machine classification network model, the feature test set and the classification prediction result data output by the optimal classification network model are input into the SHAP interpreter.

[0108] The SHAP interpreter then calculates the marginal contribution of each feature in the feature test set to the classification prediction result data, generating multiple Shapley values; these Shapley values ​​show the positive or negative contribution of each feature to the classification prediction result. The SHAP interpreter traverses all possible combinations of features and calculates the contribution value of each feature, thus providing a fair evaluation. The generated Shapley values ​​are displayed through visual graphics, such as Figure 9 The bee swarm diagram, shown here, vividly and concretely illustrates the impact of each feature on the model's decision making. By aggregating these values, we can identify the most important features for control chart pattern recognition, thereby optimizing feature selection and improving pattern recognition accuracy. SHAP not only improves the interpretability of model predictions but also provides empirical evidence to guide feature selection and model optimization.

[0109] Figure 9Each point in the graph corresponds to a sample. The horizontal axis is the Shapley value (data without units are dimensionless data). Positive and negative values ​​represent a positive or negative correlation between the feature value and the classification prediction, respectively. Larger Shapley values ​​indicate a stronger correlation. The vertical axis represents contributing features, arranged in descending order of contribution from top to bottom. Points for a particular feature are primarily concentrated in the positive direction (right side) of the Shapley value, meaning that a higher Shapley value for that feature increases the model's probability of predicting a particular class, and vice versa.

[0110] Although the nine shape features and six statistical features contribute to different control chart patterns, Figure 9 The SHAP beeswarm plot shows that the contributions of different features vary significantly. Therefore, we select the top few features with a cumulative Shapley value of 90% to construct a high-value feature subset. This is because the top 90% cumulative contribution typically captures the vast majority of important information while minimizing interference from less important features. By calculating the average absolute value of each feature's Shapley value and sorting by cumulative contribution, we can determine which features have the primary influence in the model, effectively selecting the most important features for control chart pattern recognition. This not only improves feature selection efficiency but also ensures that the selected features maximize the model's predictive power and interpretability.

[0111] In this embodiment, the nine features in the high-value feature subset that can be determined are mean, mean square value, cumulative sum statistic (cusum), kurtosis, skewness, area between the least squares line and the line segment (ass), area between the mode and the least squares line (aps), standard deviation, and slope of the least squares line.

[0112] Then, the high-value feature subset is input into the constructed graph convolutional neural network, the model is trained and the target graph convolutional neural network is obtained.

[0113] A graph convolutional neural network (GCNN) is a neural network that processes graph-structured data and can effectively capture the connections between nodes (i.e., samples). For control graph data, the similarities between samples can be constructed as graph edges. Therefore, the model input is a node vector matrix X consisting of filtered features and an adjacency matrix P representing the adjacency between control graph samples at different time series: Input = (X, P), where X is an n×d feature matrix, where n is the number of input control graph samples and d is the number of features filtered by the SHAP model. The adjacency matrix represents the similarity measure between historical and current samples in matrix form. The output of the model is different control graph patterns Y. The GCNN model performs weighted aggregation through convolution operations in two graph convolutional layers to obtain a new feature representation. Two dense layers further extract global information, and finally, an activation function is used to output the classification result Y. Figure 10 Represents the topology of a graph convolutional neural network.

[0114] The high-value feature subsets are divided into training sets and test sets, and input into the graph convolutional neural network model for training. The training process uses the backpropagation algorithm, which mainly includes two steps: forward excitation propagation and backward weight update. For forward excitation propagation, the input variables are processed by multiple graph convolution layers and then passed to the dense layer. The pattern type of the sample is output by the dense layer, and the classification results and actual results are used to calculate the loss function (error). For backward weight update, the weight of each layer is updated by the gradient descent method. When the set number of iterations is reached, the test set is used to evaluate the performance of the graph convolutional neural network, such as Figure 11 The figure shows the trend of the loss function in the embodiment as the number of iterations increases.

[0115] Can be obtained from Figure 11 As can be seen from the figure, in the early stage of the training process, the loss function of the training set decreases rapidly with the increase of the number of iterations. When the number of iterations is greater than 200, the loss function tends to a constant and no longer continues to decrease, indicating that the network has converged and is relatively stable at this time. In order to ensure the convergence of the network. The graph convolutional neural network is used to control the graph pattern classification after 500 iterations. After the model is adjusted and trained, the next step is to evaluate and verify the model performance. In order to ensure the scientificity and reliability of the evaluation, common classification evaluation indicators are used, including accuracy, recall rate and confusion matrix. Specifically, the accuracy is: The recall rate is: Among them, TP (True Positives) represents true positive examples, which are predicted to be positive examples and are actually positive examples, FP (False Positives) represents false positive examples, which are predicted to be positive examples but are actually negative examples, FN (False Negatives) represents false negative examples, which are predicted to be negative examples but are actually positive examples, and TN (True Negatives) represents true negative examples, which are predicted to be negative examples and are actually negative examples. Therefore, the accuracy rate represents the ratio of all correctly predicted samples to the total number of samples, and the recall rate represents the proportion of samples that are correctly predicted by the model as belonging to a certain category among all samples that are actually of a certain category. In this embodiment, the training set and test set of the graph convolutional neural network for control chart pattern recognition are shown in Table 2:

[0116]

[0117] And the confusion matrix of the test set is as follows Figure 12 As shown. Figure 12 It can be seen that the graph convolutional neural network in this embodiment has an average recognition accuracy of 94.3% for control chart patterns, which can realize early warning of abnormal patterns caused by different faults in the production process.

[0118] After the training is completed, the target graph convolutional neural network is obtained, which is a network with optimal parameters set.

[0119] Finally, the filtered features and trained model are used to monitor the dynamic data stream of cigarette quality. Specifically, samples of actual cigarette control charts are collected and fed into a target graph convolutional neural network for pattern recognition, identifying abnormal fluctuations in the cigarette production process.

[0120] Specifically, the longer the data's time span, the greater the lag in identifying abnormal quality characteristics. Secondly, in real-time production, when the data source is continuously and dynamically changing, sliding window technology ensures that short-term changes can be captured. Therefore, the window sequence of the measured data is set to the time span of the Monte Carlo simulation. In other words, 50 cigarette quality characteristic values ​​constitute a mean control chart sample, and a sliding window with a window movement step of 1 is used to monitor the production process in real time. In other words, for each additional cigarette quality value, the window moves back one position, and the nine features selected by the SHAP interpretable model are extracted. These features are then input into the target graph convolutional neural network for pattern classification.

[0121] Figure 13This is a real-time measurement of the dynamic data stream of cigarette quality. When the sliding window moves to the 56th data point, the recognition model issues an alarm indicating a downward trend (DT) in the cigarette manufacturing process, identifying abnormal fluctuations in the cigarette production process. By examining the corresponding control chart pattern abnormality factors, it can be determined whether the downward trend in cigarette quality is caused by changes in ambient temperature or humidity, or by equipment wear and tear that continuously affects the process. This allows for targeted solutions and other measures.

[0122] The above description is merely an illustration of the preferred embodiments disclosed in this application and the technical principles employed. Those skilled in the art should understand that the scope of protection provided by this disclosure is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this disclosure.

[0123] In addition, although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of this disclosure. Certain features described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination.

Claims

1. A cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network, characterized by: Methods include: S01. Obtaining a plurality of cigarette control chart training samples under different modes; S02, extracting training features of each cigarette control chart training sample; The training features include multiple shape features and multiple statistical features; S03, generating a feature test set based on the training features; constructing multiple classification network models, inputting the feature test set corresponding to each cigarette control chart training sample into the multiple classification network models, obtaining multiple classification prediction result data, and correspondingly obtaining ROC curve graphs; S04. Screening and obtaining an optimal classification network model according to the ROC curve graph; inputting the optimal classification network model, the feature test set, and the classification prediction result data output by the optimal classification network model into a SHAP interpreter; S05, the SHAP interpreter calculates the marginal contribution of each feature in the feature test set to the classification prediction result data, generates multiple Shapley values; and generates a high-value feature subset based on the Shapley values; S06. Inputting the high-value feature subset into the constructed graph convolutional neural network, training the model and obtaining a target graph convolutional neural network; S07. Obtain samples of the actual collected cigarette control graph and input them into the target graph convolutional neural network to perform pattern recognition and obtain cigarette quality fluctuation results.

2. The cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network according to claim 1 is characterized in that: In S01, the method for obtaining a plurality of cigarette control chart training samples under different modes is specifically: based on the control chart mode simulation formula, using the Monte Carlo method to simulate and obtain a plurality of cigarette control chart training samples under different modes.

3. The cigarette quality control graph pattern recognition method based on interpretable graph convolutional neural network according to claim 1 is characterized in that: The modes of the cigarette control chart training samples include: normal mode (NOR), cycle mode (CYC), downward trend mode (DT), upward trend mode (UT), downward step mode (DS) and upward step mode (US).

4. The cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network according to claim 1, characterized in that: The shape features include: the slope of the least squares line (slope), the average number of crossings (nc1), the number of least squares line crossings (nc2), the cycle member (cmember), the average slope of the line segment (as), the slope difference (sd) between the least squares line and the line segment representing the pattern, the area between the pattern and the mean line (apm), the area between the pattern and its least squares line (aps), and the area between the least squares line and the line segment (ass).

5. The cigarette quality control graph pattern recognition method based on interpretable graph convolutional neural network according to claim 1 is characterized in that: The statistical features include mean, cumulative sum statistic (cusum), skewness, kurtosis, mean square value (mean square value) and standard deviation (standard deviation).

6. The cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network according to claim 1, characterized in that: In S03, the multiple classification network models constructed are: decision tree model, random forest model, logistic regression model, support vector machine model, XGBoost model and GBDT model.

7. The cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network according to claim 6, characterized in that: In S05, the specific method of generating a high-value feature subset based on the Shapley value is: calculating the average absolute value of each Shapley value and sorting them according to the cumulative proportion, and selecting features with the top 90% cumulative proportion of Shapley values ​​to form a high-value feature subset.

8. The cigarette quality control graph pattern recognition method based on an interpretable graph convolutional neural network according to claim 7, characterized in that: In S06, the constructed graph convolutional neural network is trained using a back-propagation algorithm to obtain the target graph convolutional neural network.

Images