System for monitoring and adjusting quality of pickled vegetables in old jar

Abstract

The invention discloses a quality detection and adjustment system for pickled vegetables in an old jar. Through application of the technical scheme of the invention, on one hand, the BP neural network can be used for training data to autonomously learn and adapt to various complex nonlinear relationships, the quality characteristics of the pickled vegetables can be identified and evaluated more accurately, large-scale data can be processed, higher prediction precision can be provided, producers can be helped to better control the quality of the pickled vegetables, and the production efficiency is improved. And personal errors are reduced. On the other hand, the quality of the pickled vegetables is influenced by various factors such as moisture content, nitrite content, microbial indicators, heavy metal content, food additive content and nutritional ingredients, the BP neural network can process various characteristic data at the same time, the quality is comprehensively evaluated, the BP neural network can quickly feed back quality information through real-time data input, and the quality of the pickled vegetables is improved. Immediate adjustment in the production process is facilitated, and the production efficiency is improved.

Description

Technical Field

[0001] The invention relates to the field of food quality monitoring and adjustment, and in particular to a quality detection and adjustment system for old-style pickled vegetables. Background Art

[0002] As people's living standards gradually improve, the public gradually pays more attention to the quality and safety of food. Laotan pickles are the most common food, and because of its unique taste and rich nutrition, the demand and sales of Laotan pickles are increasing.

[0003] Monitoring and adjusting the quality of old pickles can improve product quality, reduce food safety risks, strengthen food safety management standards, and improve industry management levels. However, most of the current methods for detecting and adjusting the quality of old pickles are in the manual observation stage. Manual monitoring and adjustment have large errors, low efficiency, and cannot process multiple feature data. The scope and quantity of improving the product quality of old pickles are small.

[0004] Therefore, it is necessary to design a quality detection and adjustment system for old pickles to solve the problems of low efficiency, large errors, inability to adjust in time, inability to process multiple feature data, and inability to be implemented on a large scale in traditional methods of monitoring and adjusting the quality of old pickles. Summary of the invention

[0005] In view of this, the present invention proposes a quality detection and adjustment system for old-style pickles, aiming to solve the problems of low efficiency, large error, inaccurate prediction results, inability to adjust in time, inability to process multiple feature data, and inability to be implemented on a large scale in traditional methods of monitoring and adjusting the quality of old-style pickles.

[0006] In one aspect, the present invention provides a quality detection and adjustment system for old pickles, comprising:

[0007] A data collection module is used to collect historical core data and real-time core data of Laotan pickles; the core data includes influencing factor data and predicted target data;

[0008] A data processing module, used for cleaning the historical core data and real-time core data of the old pickles, and removing abnormal values ​​and missing values ​​in the data;

[0009] A prediction module, used for inputting the historical core data into the Laotan pickle quality prediction model, so as to train, verify, adjust and test the Laotan pickle quality prediction model, so that the prediction performance of the Laotan pickle quality prediction model reaches the best;

[0010] The adjustment module inputs the real-time core data into the Laotan pickles quality prediction model to obtain the prediction results of the Laotan pickles quality. Based on the prediction results, the real-time core data is adjusted in real time to simulate the influence of different core influencing factors on the quality of Laotan pickles.

[0011] Furthermore, the historical core data is divided into training data, verification data and test data.

[0012] The training data is used to train the quality prediction model of Laotan pickles;

[0013] The verification data is used to adjust the parameters of the old pickle quality prediction model through iterative training;

[0014] The test data is used to evaluate the prediction performance of the old-style pickled vegetable quality prediction model.

[0015] Furthermore, the prediction module includes a BP neural network structure design unit, a weight initialization unit, a forward propagation unit, a weight update unit, an iterative training unit and an evaluation optimization unit;

[0016] The BP neural network structure design unit is used to design a BP neural network structure for the Laotan pickles quality prediction model based on the historical core data;

[0017] The weight initialization unit is used to assign initial values ​​to connection weights in the neural network;

[0018] The forward propagation unit is used to input the historical core data into the Laotan pickles quality prediction model to obtain the prediction result of the Laotan pickles quality index;

[0019] The weight updating unit is used to update the weights and biases in the BP neural network;

[0020] The iterative training unit is used to adjust the hyperparameters;

[0021] The evaluation and optimization unit is used to evaluate the quality prediction model of old-style pickles and take a series of measures to optimize the quality prediction model of old-style pickles.

[0022] Furthermore, the BP neural network structure design unit includes determining the number of layers, determining the number of nodes and selecting an activation function;

[0023] The BP neural network structure consists of three layers, namely, input layer, hidden layer and output layer;

[0024] The number of nodes in the input layer is the same as the number of types of influencing factor data in the historical core data;

[0025] The number of nodes in the hidden layer is determined by the hidden layer node formula, which is expressed as:

[0026] N y =N r +N c +a

[0027] Among them, N y Represents the number of nodes in the hidden layer, N r Represents the number of nodes in the input layer, N c represents the number of nodes in the output layer, and a represents a constant between 1 and 10;

[0028] The number of nodes in the output layer is the same as the number of types of predicted target data in the historical core data;

[0029] The activation function is used to introduce nonlinearity, which enables the neural network to fit complex functions.

[0030] Furthermore, the initialization method of the weight initialization unit includes a random initialization method and a special initialization method;

[0031] The random initialization method uses normal distribution or uniform distribution to initialize weights;

[0032] The special initialization method uses the characteristics of the network structure and activation function to initialize the weights.

[0033] Furthermore, the forward propagation unit includes input data, layer-by-layer calculation and output results;

[0034] The input data is the training data in the historical core data inputted into the input layer of the BP neural network structure;

[0035] The layer-by-layer calculation is to calculate the output value of each neuron layer by layer in the hidden layer of the BP neural network structure according to the network structure and weight;

[0036] The output result is a prediction result outputted by the output layer of the BP neural network structure;

[0037] The input data, the layer-by-layer calculation and the output result are expressed by a mathematical relationship formula, and the mathematical relationship formula is expressed as follows:

[0038]

[0039] Among them, F y Represents the output of the hidden layer, W ij represents the weight of the hidden layer, W jk represents the weight of the output layer, F crepresents the output of the output layer, a j represents the bias of the hidden layer, b k represents the bias of the output layer, G represents the activation function, and X represents the data input by the input layer.

[0040] Further, the weight updating unit includes calculating system errors and updating weights;

[0041] The calculation system error is expressed by an error formula, which is expressed as:

[0042]

[0043] Where E represents the system error, F q Indicates the expected output of the system, F c Indicates the actual output;

[0044] The updated weight is expressed by a weight update formula, which is expressed as:

[0045] e k =F q -F c

[0046]

[0047] W jk =W jk +βF y e k

[0048]

[0049] b k =b k +βe k

[0050] Among them, β is the learning rate.

[0051] Furthermore, the iterative training unit utilizes the verification data to iteratively train the BP neural network by repeatedly performing the forward propagation unit and the weight updating unit.

[0052] Furthermore, the evaluation and optimization unit evaluates the quality prediction model of old-style pickles through the test data, and takes measures to optimize the quality prediction model of old-style pickles.

[0053] Furthermore, the adjustment module inputs the real-time core data into the quality prediction model of the old-style pickled vegetables, and adjusts the influencing factor data according to the obtained prediction results, so that the quality of the old-style pickled vegetables reaches the optimal level.

[0054] Compared with the prior art, the beneficial effects of the present invention are: the quality prediction model of Laotan pickles can accurately predict the quality of Laotan pickles and make timely and effective adjustments, can process multiple feature data at the same time, can quickly feedback quality information, help instant adjustments in the production process, and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0056] Figure 1 The present invention provides a functional block diagram of a quality detection and adjustment system for traditional pickled vegetables. DETAILED DESCRIPTION

[0057] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present application.

[0058] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0059] The following description of at least one exemplary embodiment is merely illustrative in nature and is not intended to limit the present application, its application, or uses.

[0060] Technologies, systems, and devices known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, systems, and devices should be considered part of the specification.

[0061] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0062] In addition, the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0063] It should be noted that all directional indications in the embodiments of the present application (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0064] Combine the following Figure 1 To describe the quality detection and adjustment system for old pickles according to the exemplary embodiment of the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principle of the present application, and the embodiments of the present application are not limited in this regard. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0065] Furthermore, the present application proposes a quality detection and adjustment system for old-style pickles.

[0066] Figure 1 The following schematically shows a workflow diagram of a system for detecting and adjusting the quality of old pickles according to an embodiment of the present application. Figure 1 As shown, including:

[0067] 100, a data collection module, is used to collect historical core data and real-time core data of Laotan pickles.

[0068] 200, a data processing module, is used to clean the historical core data and real-time core data of Laotan pickles, and remove abnormal values ​​and missing values ​​in the data.

[0069] 300, a prediction module, is used to input the historical core data into the old pickle quality prediction model, so as to train, verify, adjust and test the old pickle quality prediction model, so that the prediction performance of the old pickle quality prediction model reaches the best.

[0070] 400, the adjustment module inputs the real-time core data into the quality prediction model of the old-style pickles to obtain the prediction result of the quality of the old-style pickles, and adjusts the real-time core data in real time based on the prediction result to simulate the influence of different core influencing factor contents on the quality of the old-style pickles.

[0071] Specifically, the core data includes influencing factor data and predicted target data. The influencing factor data are factors that affect the quality of old-style pickles. In this embodiment, the influencing factor data include moisture content, nitrite content, microbial indicators, acid value and peroxide value, pH value and heavy metal content. The predicted target data is the quality index of old-style pickles.

[0072] It is understandable that the more types of influencing factor data there are, the more accurate the prediction result will be. The multiple influencing factor data provided in this embodiment can greatly improve the accuracy of the quality prediction of Laotan pickles.

[0073] Specifically, historical core data are divided into training data, verification data and test data. The training data is used to train the quality prediction model of Laotan pickles, the verification data is used to adjust the parameters of the quality prediction model of Laotan pickles through iterative training, and the test data is used to evaluate the prediction performance of the quality prediction model of Laotan pickles. The training data accounts for 70% to 80% of the historical core data, the verification data accounts for 10% to 15% of the historical core data, and the test data accounts for 10% to 15% of the historical core data.

[0074] It is understandable that dividing the historical core data into training data, validation data, and test data can improve the generalization ability of the model, avoid overfitting, and objectively evaluate the performance of the model. The model can learn the inherent laws of the data by training the model with training data. The validation data is used to monitor the model error during the training process, helping to stop the training in time to avoid overfitting, thereby ensuring the model's prediction ability on unseen data. The test data is used to evaluate the generalization ability of the model and ensure that the model can accurately predict unseen data. The validation data continuously tests the model's error during the training process to avoid overfitting by identifying the error inflection point. When the error of the validation data reaches an inflection point, the training is stopped, which ensures that the model will not overfit on the training data, thereby improving the performance of the model in practical applications. The test data is used to test the model's error after the neural network training is completed, which helps to objectively evaluate the performance of the model. By separating the training data, validation data, and test data, the generalization ability of the model can be more accurately evaluated, ensuring that the prediction results of the model have high credibility and accuracy.

[0075] Specifically, the prediction module includes a BP neural network structure design unit, a weight initialization unit, a forward propagation unit, a weight update unit, an iterative training unit, and an evaluation and optimization unit.

[0076] Specifically, the BP neural network structure design unit is used to design the BP neural network structure for the Laotan pickles quality prediction model based on the historical core data. The BP neural network structure has three layers, namely, an input layer, a hidden layer and an output layer.

[0077] Specifically, the input layer is used to receive external input data and serves as an interface for the neural network to interact with the outside world. The number of neurons in the input layer depends on the number of features of the input data; the hidden layer is used to perform nonlinear transformation and feature extraction on the input data, and is the core part of the neural network. The hidden layer can learn the complex mapping relationship between input and output through multiple layers and multiple neurons, thereby improving the complexity and learning ability of the network; the output layer is used to generate the final prediction result, and is designed according to the specific goals of the problem (such as classification, regression, etc.). The number of neurons in the output layer depends on the output requirements of the problem.

[0078] Specifically, the BP neural network structure design unit is used to determine the number of layers, the number of nodes and the selection of the activation function. The number of nodes in the input layer is the same as the number of types of influencing factor data in the historical core data; the number of nodes in the hidden layer is determined by the hidden layer node formula, which is expressed as:

[0079] N y =N r +N c +a

[0080] Among them, N y Represents the number of nodes in the hidden layer, N r Indicates the number of nodes in the input layer, N c represents the number of nodes in the output layer, and a represents a constant between 1 and 10; the number of nodes in the output layer is the same as the number of types of predicted target data in the historical core data; the activation function is used to introduce nonlinearity, which enables the neural network to fit complex functions.

[0081] It is understandable that the choice of the number of layers depends on the complexity of the specific problem and the amount of data. The more hidden layers there are, the stronger the predictive ability of the model.

[0082] In this embodiment, the excitation function is a Sigmoid function. It can be understood that the Sigmoid function can map the input value to a range between 0 and 1. This nonlinear mapping capability enables the neural network to learn and represent complex nonlinear relationships, thereby better fitting the data.

[0083] Specifically, the weight initialization unit is used to assign initial values ​​to the connection weights in the neural network. Before training the model, it is necessary to assign initial values ​​to the connection weights in the neural network. The quality of weight initialization has a great impact on the convergence speed and prediction performance of the model.

[0084] In this embodiment, the normal distribution is selected to initialize the weights. It can be understood that the normal distribution can prevent all neurons from having the same output in the early stage of training, thereby accelerating convergence.

[0085] Specifically, the forward propagation unit includes input data, layer-by-layer calculations, and output results;

[0086] The input data is the training data in the historical core data input into the input layer of the BP neural network structure;

[0087] Layer-by-layer calculation is to calculate the output value of each neuron layer by layer in the hidden layer of the BP neural network structure according to the network structure and weight;

[0088] The output result is the prediction result outputted in the output layer of the BP neural network structure;

[0089] The input data, layer-by-layer calculations, and output results are expressed using mathematical formulas. The mathematical formulas are:

[0090]

[0091] Among them, F y Represents the output of the hidden layer, W ij represents the weight of the hidden layer, W jk represents the weight of the output layer, F c represents the output of the output layer, a j represents the bias of the hidden layer, b k represents the bias of the output layer, G represents the activation function, and X represents the data input to the input layer.

[0092] It is understandable that the main advantage of forward propagation is that it can efficiently process and transmit information while keeping the calculation simple and efficient. This propagation method enables BP neural network to perform well in dealing with complex problems and large-scale data, especially in solving nonlinear problems. The simplicity of forward propagation makes BP neural network easy to understand and implement, and is suitable for many types of neural network structures, such as multi-layer perceptron and convolutional neural network. This structural advantage further enhances the flexibility and effect of BP neural network in practical applications.

[0093] Specifically, the weight updating unit includes calculating the system error and updating the weight;

[0094] The calculation system error is expressed by the error formula, which is expressed as:

[0095]

[0096] Where E represents the system error, F q Indicates the expected output of the system, F c Indicates the actual output;

[0097] The updated weight is expressed by the weight update formula, which is expressed as:

[0098] e k =F q -F c

[0099]

[0100] W jk =W jk +βF y e k

[0101]

[0102] b k =b k +βe k

[0103] Among them, β is the learning rate.

[0104] 1. Specifically, the adjustment of weights and biases is completed by the back propagation of the BP neural network. Starting from the output layer, the gradient of each weight and bias is calculated layer by layer using the chain rule according to the gradient of the loss function; for each weight and bias, the gradients from all training samples are accumulated (or averaged) to obtain the final gradient value; in order to avoid the gradient explosion problem, it is sometimes necessary to clip the gradient value, that is, when the gradient value exceeds a certain threshold, it is truncated to the threshold.

[0105] 2. In this embodiment, the algorithm used for weight update is the gradient descent method. It can be understood that the gradient descent method can be applied to most machine learning models, whether linear or nonlinear problems. Through the iterative process, the gradient descent method can automatically adjust the learning rate, thereby improving the convergence speed. For large-scale data sets, the gradient descent method can effectively optimize parameters. The gradient descent method gradually approaches the minimum value by continuously updating the model parameters along the negative gradient direction of the loss function.

[0106] 3. Specifically, the iterative training unit uses the validation data to iteratively train the BP neural network by repeatedly performing the forward propagation unit and the weight update unit. During the training process, it is necessary to monitor the loss changes of the validation data and the prediction performance of the model. This helps to promptly discover problems such as overfitting or underfitting and take appropriate measures to make adjustments. Iterative training can adjust hyperparameters: hyperparameters include learning rate, batch size, number of hidden layer nodes, number of iterations, etc. During the training process, these hyperparameters may need to be adjusted according to the performance of the model to obtain better prediction performance.

[0107] 4. Specifically, the evaluation and optimization unit evaluates the quality prediction model of Laotan pickles through test data, and takes measures to optimize the quality prediction model of Laotan pickles. The model is evaluated using test data, and the prediction accuracy is calculated to measure the performance of the model. According to the evaluation results, a series of measures can be taken to optimize the model. For example, adjust the network structure, increase the amount of data, use more complex features, try different optimization algorithms, etc.

[0108] 5. Specifically, the adjustment module inputs the real-time core data into the quality prediction model of Laotan pickles, and adjusts the influencing factor data according to the prediction results to achieve the best quality of Laotan pickles.

[0109] It can be understood that this embodiment can accurately predict the quality of Laotan pickles and make timely and effective adjustments through the Laotan pickle quality prediction model, can process multiple feature data at the same time, can quickly feedback quality information, help instant adjustments in the production process, and improve production efficiency.

[0110] It should be noted that:

[0111] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known structures and technologies are not shown in detail so as not to obscure the understanding of this description.

[0112] Similarly, it should be understood that in order to streamline the present application and aid in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed system should not be interpreted as reflecting the following schematic diagram: the claimed application requires more features than those explicitly recited in each claim.

[0113] Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment.The claims following the Detailed Description are thus hereby expressly incorporated into this Detailed Description, with each claim standing on its own as a separate embodiment of this application.

[0114] Furthermore, those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present application and to form different embodiments.

[0115] For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0116] 6. The above description is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A quality monitoring and adjustment system for old pickles, characterized in that: The system includes a data acquisition module, a data processing module, a prediction module, and an adjustment module; The data collection module is used to collect historical core data and real-time core data of Laotan pickles; the core data includes influencing factor data and predicted target data; The data processing module is used to clean the historical core data and real-time core data of the old pickles to remove abnormal values ​​and missing values ​​in the data; The prediction module is used to input the historical core data into the Laotan pickles quality prediction model to train, verify, adjust and test the Laotan pickles quality prediction model so that the prediction performance of the Laotan pickles quality prediction model reaches the best; The adjustment module inputs the real-time core data into the Laotan pickles quality prediction model to obtain the prediction results of the Laotan pickles quality. Based on the prediction results, the real-time core data is adjusted in real time to simulate the influence of different core influencing factor contents on the Laotan pickles quality.

2. The quality monitoring and adjustment system for old pickles according to claim 1 is characterized in that: The historical core data is divided into training data, verification data and test data The training data is used to train the quality prediction model of Laotan pickles; The verification data is used to adjust the parameters of the old pickle quality prediction model through iterative training; The test data is used to evaluate the prediction performance of the old-style pickled vegetable quality prediction model.

3. The quality monitoring and adjustment system for traditional pickles according to claim 1 is characterized in that: The prediction module includes a BP neural network structure design unit, a weight initialization unit, a forward propagation unit, a weight update unit, an iterative training unit and an evaluation optimization unit; The BP neural network structure design unit is used to design a BP neural network structure for the Laotan pickles quality prediction model based on the historical core data; The weight initialization unit is used to assign initial values ​​to connection weights in the neural network; The forward propagation unit is used to input the historical core data into the Laotan pickles quality prediction model to obtain the prediction result of the Laotan pickles quality index; The weight updating unit is used to update the weights and biases in the BP neural network; The iterative training unit is used to adjust the hyperparameters; The evaluation and optimization unit is used to evaluate the quality prediction model of old-style pickles and take a series of measures to optimize the quality prediction model of old-style pickles.

4. The quality monitoring and adjustment system for traditional pickles according to claim 3 is characterized in that: The BP neural network structure design unit includes determining the number of layers, determining the number of nodes and selecting an activation function; The BP neural network structure consists of three layers, namely, input layer, hidden layer and output layer; The number of nodes in the input layer is the same as the number of types of influencing factor data in the historical core data; The number of nodes in the hidden layer is determined by the hidden layer node formula, which is expressed as: N y =N r +N c +a Among them, N y Represents the number of nodes in the hidden layer, N r Represents the number of nodes in the input layer, N c represents the number of nodes in the output layer, and a represents a constant between 1 and 10; The number of nodes in the output layer is the same as the number of types of predicted target data in the historical core data; The activation function is used to introduce nonlinearity, which enables the neural network to fit complex functions.

5. The quality monitoring and adjustment system for traditional pickles according to claim 3 is characterized in that: The initialization method of the weight initialization unit includes a random initialization method and a special initialization method; The random initialization method uses normal distribution or uniform distribution to initialize weights; The special initialization method uses the characteristics of the network structure and activation function to initialize the weights.

6. A quality monitoring and adjustment system for old pickles according to any one of claims 1 to 5, characterized in that: The forward propagation unit includes input data, layer-by-layer calculation and output results; The input data is the training data in the historical core data inputted into the input layer of the BP neural network structure; The layer-by-layer calculation is to calculate the output value of each neuron layer by layer in the hidden layer of the BP neural network structure according to the network structure and weight; The output result is a prediction result outputted by the output layer of the BP neural network structure; The input data, the layer-by-layer calculation and the output result are expressed by a mathematical relationship formula, and the mathematical relationship formula is expressed as follows: Among them, F y Represents the output of the hidden layer, W ij represents the weight of the hidden layer, W jk represents the weight of the output layer, F c represents the output of the output layer, a j represents the bias of the hidden layer, b k represents the bias of the output layer, G represents the activation function, and X represents the data input by the input layer.

7. The quality monitoring and adjustment system for traditional pickles according to claim 6 is characterized in that: The weight updating unit includes calculating system error and updating weight; The calculation system error is expressed by an error formula, which is expressed as: Where E represents the system error, F q Indicates the expected output of the system, F c Indicates the actual output; The updated weight is expressed by a weight update formula, which is expressed as: e k =F q -F c IN jk =In jk +βF y and k b k =b k +βe k Among them, β is the learning rate.

8. The quality monitoring and adjustment system for traditional pickles according to claim 7 is characterized in that: The iterative training unit utilizes the verification data to iteratively train the BP neural network by repeatedly performing the forward propagation unit and the weight updating unit.

9. The system for monitoring and adjusting the quality of old pickles according to any one of claim 8, characterized in that: The evaluation and optimization unit evaluates the quality prediction model of old-style pickles through the test data, and takes measures to optimize the quality prediction model of old-style pickles.

10. The quality monitoring and adjustment system for traditional pickles according to claim 1 is characterized in that: The adjustment module inputs the real-time core data into the quality prediction model of the old-style pickles, and adjusts the influencing factor data according to the obtained prediction results, so that the quality of the old-style pickles is optimized.

Images