Food quality prediction system based on big data and fuzzy mathematics

Abstract

The invention discloses a food quality prediction system based on big data and fuzzy mathematics, which comprises a data acquisition module, a data processing module, a comprehensive deviation rate generation module and a quality prediction model module, and relates to the technical field of food safety intelligent detection. The comprehensive deviation rate generation module effectively captures fuzziness of food component contents through a fuzzy mathematical theory, differential influences are formed by taking final weights of all the components in the food components as differential weights, and interaction coefficients of all the components in the food components form collaborative deviation risks among the components. Therefore, the calculation accuracy of the comprehensive deviation rate is improved, the problems of insufficient uncertainty processing, fuzzy key factor identification and collaborative risk neglect in food component quality evaluation are systematically solved, and the comprehensiveness and accuracy of quality risk prediction in a subsequent quality prediction model are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of food safety intelligent detection, and in particular to a food quality prediction system based on big data and fuzzy mathematics. Background Art

[0002] Food quality testing refers to the process of analyzing, evaluating, and verifying food ingredients, safety, and quality characteristics through scientific methods and techniques to determine whether it complies with relevant standards, regulations, and expected quality requirements. This process encompasses comprehensive monitoring of food raw materials, production and processing, finished products, and distribution.

[0003] Currently, existing food quality inspections usually use fixed thresholds to determine whether the ingredient content is qualified, ignoring the fuzzy interval characteristics of the standard content, resulting in insufficient uncertainty processing; when allocating weights for each ingredient in food, static empirical values ​​are often used, and the importance is not adjusted based on the actual deviation of the ingredients, resulting in fuzzy identification of key factors; in addition, traditional food quality inspection models are difficult to capture multi-factor nonlinear correlations, especially in complex production environments, where equipment operation fluctuations and ingredient synergistic / antagonistic effects are easily overlooked, leading to the problem of ignoring synergistic risks, which in turn leads to insufficient accuracy in food quality predictions. Summary of the Invention

[0004] In order to solve the technical problems existing in the background technology, the present invention provides a food quality prediction system based on big data and fuzzy mathematics.

[0005] The present invention proposes a food quality prediction system based on big data and fuzzy mathematics, comprising:

[0006] Data acquisition module: used to collect multi-source data of the entire food production process;

[0007] Multi-source data includes food ingredient standard content, food ingredient actual content, food ingredients, production environment parameters, production equipment operation data, and food ingredient testing data;

[0008] Data processing module: used to clean and classify the multi-source data collected in the data acquisition module;

[0009] Comprehensive deviation rate generation module: obtains multi-source data after cleaning and classification in the data processing module;

[0010] The fuzzy set of food ingredient standard content is constructed by using fuzzy mathematics theory, and the value range of the fuzzy set is defined by the trapezoidal membership function to form the fuzzy standard content.

[0011] The difference between the actual content of food ingredients and the fuzzy standard content is quantified by the Hamming fuzzy distance formula to obtain the ingredient fuzzy deviation rate;

[0012] The importance weights of each component in the food ingredients are represented by triangular fuzzy numbers, and the final weights of each component in the food ingredients are determined by the fuzzy comprehensive evaluation method. The weighted deviation of the food ingredients is obtained by combining the component fuzzy deviation rate.

[0013] The interaction coefficient of each component in the food ingredients is determined by the component fuzzy deviation rate, and the comprehensive deviation rate is obtained by combining the weighted deviation of the food ingredients.

[0014] Quality prediction model module: A quality risk prediction model is constructed through the comprehensive deviation rate obtained by the comprehensive deviation rate generation module and the production environment parameters. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability in the future unit time.

[0015] Preferably, in the comprehensive deviation rate generation module, the value range of the fuzzy set is defined by the trapezoidal membership function as follows:

[0016]

[0017] where a, b, c, d are the boundary parameters of the fuzzy interval, and 0 < a < b < c < d, x is the actual content of the food ingredient, and μ(x) is the membership degree.

[0018] Preferably, in the comprehensive deviation rate generation module, the weighted deviation of the food ingredients obtained by combining the final weights of each component in the food ingredients and the component fuzzy deviation rate is generated in the following manner:

[0019] Let the number of types of food ingredients be n, i = 1, 2, 3,..., n, w ,

[0022] ,

[0021] , i , i ,

[0020] , i ,

[0019] ,

[0018] ,

[0024] ,

[0017] ,

[0023] , , , <00000​​​​​​​​​​​​​​​​​​​​​​​​​Assume that the fuzzy deviation rate of the jth food ingredient is d j , j≠i;

[0025] Let the interaction coefficient between the i-th food component and the j-th food component be c ij , then c ij The calculation formula is: ij =d i ·d j , if the deviation rate of food ingredient d j is larger, and the deviation rate d of the jth food ingredient j is also larger, and the product of the two is c ij It will increase significantly, indicating that when the deviations of these two components exist at the same time, the comprehensive impact of their interaction on food quality is more prominent.

[0026] Preferably, the interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and combined with the weighted deviation to obtain the comprehensive deviation rate, as follows:

[0027] Assume that the comprehensive deviation rate is S and the weighted deviation of the jth food component is B j , the calculation formula of S is:

[0028]

[0029] Direct accumulation of weighted deviations for all food components;

[0030] The interaction between any two different components is expressed by the interaction coefficient c ij Coupling effects that amplify or reduce deviations.

[0031] Preferably, in the quality prediction model module, a quality risk prediction model is constructed as follows:

[0032] Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient testing data obtained by the comprehensive deviation rate generation module, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rules;

[0033] A quality risk prediction model is constructed using the random forest algorithm based on the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient testing data, and the comprehensive deviation rate per unit time in the future;

[0034] In the quality risk prediction model:

[0035] When the input is the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient testing data, the output of the quality risk prediction model is the current quality risk probability;

[0036] When the input is the comprehensive deviation rate per unit time in the future, production environment parameters, production equipment operation data, and food ingredient testing data, the output of the quality risk prediction model is the quality risk probability per unit time in the future.

[0037] Preferably, in the quality prediction model module, the current quality risk probability output by the quality risk prediction model is set as P1, and the quality risk probability per unit time in the future is set as P2;

[0038] When P1 is greater than or equal to the set threshold, the alarm signal is triggered;

[0039] When P2 is greater than or equal to the set threshold, an early warning signal is triggered.

[0040] A food quality prediction method based on big data and fuzzy mathematics includes the following steps:

[0041] S1. Collect multi-source data from the entire food production process;

[0042] Multi-source data includes food ingredient standard content, food ingredient actual content, food ingredients, production environment parameters, production equipment operation data, and food ingredient testing data;

[0043] S2, cleans and classifies the multi-source data collected in S1;

[0044] S3, obtaining the multi-source data after cleaning and classification in S2, constructing a fuzzy set of food ingredient standard content using fuzzy mathematics theory, defining the value range of the fuzzy set through a trapezoidal membership function, and forming a fuzzy standard content;

[0045] The difference between the actual content of food ingredients and the fuzzy standard content is quantified by the Hamming fuzzy distance formula to obtain the ingredient fuzzy deviation rate;

[0046] The importance weight of each component in food composition is expressed by triangular fuzzy number, the final weight of each component in food composition is determined by fuzzy comprehensive evaluation method, and the weighted deviation of food composition is obtained by combining the fuzzy deviation rate of the component.

[0047] The interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and the comprehensive deviation rate is obtained by combining the weighted deviation of the food components;

[0048] S4. Based on the comprehensive deviation rate obtained in S3, production environment parameters, production equipment operation data, and food ingredient testing data, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rule;

[0049] Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient detection data and the comprehensive deviation rate per unit time in the future, a quality risk prediction model is constructed using the random forest algorithm. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability per unit time in the future.

[0050] The food quality prediction system proposed in the present invention based on big data and fuzzy mathematics has the following beneficial technical effects:

[0051] 1. The comprehensive deviation rate generated by the comprehensive deviation rate generation module integrates the standard content of food ingredients, the actual content of food ingredients, and the food composition. It not only includes the weighted deviation of individual food ingredients, but also determines the interaction coefficient of each ingredient in the food composition through the ingredient fuzzy deviation rate, forming a quantitative indicator covering the overall deviation degree of the food ingredient system. It can accurately capture the synergistic enhancement or antagonistic offset effects between ingredients, thereby more comprehensively assessing food quality risks and effectively solving the defect of traditional assessment methods that ignore the interaction between ingredients.

[0052] The comprehensive deviation rate generation module effectively captures the ambiguity of food ingredient content through fuzzy mathematics theory, forms differentiated influences through the final weights of each ingredient in the food ingredients as differentiated weights, and forms the collaborative deviation risk between ingredients through the interaction coefficients of each ingredient in the food ingredients, thereby improving the calculation accuracy of the comprehensive deviation rate. It systematically solves the problems of insufficient uncertainty processing, vague identification of key factors, and neglect of collaborative risks in food ingredient quality assessment, and improves the comprehensiveness and accuracy of quality risk prediction in subsequent quality prediction models.

[0053] 2. Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient detection data obtained by the comprehensive deviation rate generation module, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rules, and the random forest algorithm is used to construct a quality risk prediction model, thereby improving the accuracy and reliability of quality risk prediction.

[0054] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a principle block diagram of the system of the present invention;

[0056] Figure 2 A flow chart showing the comprehensive deviation rate generation of the comprehensive deviation rate generation module of the present invention;

[0057] Figure 3 Flowchart of the method of the present invention. DETAILED DESCRIPTION

[0058] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention, and are not to be construed as limiting the present invention.

[0059] like Figure 1-Figure 2 A food quality prediction system based on big data and fuzzy mathematics is shown, comprising:

[0060] Data acquisition module: used to collect multi-source data of the entire food production process;

[0061] Multi-source data includes food ingredient standard content, food ingredient actual content, food ingredients, production environment parameters, production equipment operation data, and food ingredient testing data;

[0062] Data processing module: used to clean and classify the multi-source data collected in the data acquisition module;

[0063] Comprehensive deviation rate generation module: obtains multi-source data after cleaning and classification in the data processing module;

[0064] The fuzzy set of food ingredient standard content is constructed by using fuzzy mathematics theory, and the value range of the fuzzy set is defined by the trapezoidal membership function to form the fuzzy standard content.

[0065] The difference between the actual content of food ingredients and the fuzzy standard content is quantified by the Hamming fuzzy distance formula to obtain the ingredient fuzzy deviation rate;

[0066] The importance weight of each component in the food composition is expressed by triangular fuzzy numbers. The final weight of each component in the food composition is determined by fuzzy comprehensive evaluation method, thereby forming a differentiated weight. The final weight of each component in the food composition is combined with the component fuzzy deviation rate to obtain the weighted deviation of the food composition.

[0067] The interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and the comprehensive deviation rate is obtained by combining the weighted deviation of the food components;

[0068] In this application, by setting up a comprehensive deviation rate generation module, a fuzzy set of standard food ingredient contents is constructed by introducing the theory of fuzzy mathematics. The value range of the fuzzy set is defined by the trapezoidal membership function to form a fuzzy standard content, which can adapt to the quality standard differences of different food categories and better fit the natural fluctuation characteristics of ingredient contents in food production. By using the Hamming fuzzy distance formula to quantify the difference between the actual content of food ingredients and the fuzzy standard content, the ingredient fuzzy deviation rate is obtained. Compared with the traditional Euclidean distance or absolute error, the Hamming fuzzy distance can more accurately capture the changes in the actual content of food ingredients and better conform to the actual law of quality fluctuations. The importance weights of each ingredient in food ingredients are represented by triangular fuzzy numbers, and the final weights of each ingredient in food ingredients are determined by the fuzzy comprehensive evaluation method, solving the problem that traditional fixed weights are difficult to reflect complex quality influencing factors and avoiding the evaluation distortion where the deviation of minor ingredients masks the core risks. The interaction coefficients of each ingredient in food ingredients are determined by the ingredient fuzzy deviation rate to quantify the non-linear interaction of deviations between ingredients and make the interaction risks explicit. When two ingredients deviate from the standard simultaneously, the interaction term will amplify the comprehensive deviation rate, truly reflecting the risk superposition effect of complex ingredient systems and avoiding underestimating the quality hidden dangers caused by multi-factor coupling.

[0069] Quality prediction model module: Based on the comprehensive deviation rate obtained from the comprehensive deviation rate generation module and the multi-source data that have been cleaned and classified in the data processing module, a quality risk prediction model is constructed. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability for the next unit of time.

[0070] Furthermore, in the comprehensive deviation rate generation module, the value range of the fuzzy set is defined by the trapezoidal membership function as follows:

[0071]

[0072] where a, b, c, d are the boundary parameters of the fuzzy interval, and 0 < a < b < c < d, x is the actual content of food ingredients, and μ(x) is the membership degree.

[0073] Furthermore, in the comprehensive deviation rate generation module, the weighted deviation of food ingredients is generated by combining the final weight of each ingredient in food ingredients with the ingredient fuzzy deviation rate in the following way:

[0074] Let the number of types of food ingredients be n, i = 1, 2, 3,..., n, w i be the final weight of the i-th type of food ingredient; it reflects the importance of this ingredient in the food quality system. The higher the weight, the greater the impact of this ingredient on food quality.

[0075] Let the ingredient fuzzy deviation rate of the i-th type of food ingredient be d i ;

[0076] The weighted deviation B of the i-th food component i The calculation formula is:

[0077] B i =w i ·d i .

[0078] When the weight wi of a food component is high and the fuzzy deviation rate di is also large, Bi will increase significantly, thus occupying a more critical position in the comprehensive evaluation;

[0079] By integrating multi-source information such as food ingredient data, production environment parameters, and production equipment operating data, and utilizing intelligent data cleaning technology, the system has built a comprehensive data foundation covering the entire production process, completely overcoming the limitations of traditional methods that rely on a single data source. Secondly, through the innovative application of fuzzy mathematics theory, trapezoidal membership functions are used to flexibly define the reasonable fluctuation range of the standard content of ingredients. The Hamming fuzzy distance algorithm is used to accurately calculate the degree of deviation between the actual content and the standard value. Combined with a dynamically adjusted triangular fuzzy number weight distribution mechanism, the system can accurately quantify quality uncertainty factors that are difficult to handle with traditional methods.

[0080] Furthermore, in the comprehensive deviation rate generation module, the interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, as follows:

[0081] Assume that the fuzzy deviation rate of the jth food ingredient is d j , j≠i, indicating that j represents other components different from the i-th component. j It is obtained by quantifying the difference between the actual content of the jth component and the fuzzy standard content through methods such as the Hamming fuzzy distance formula, which intuitively reflects the degree to which its content deviates from the standard;

[0082] Let the interaction coefficient between the i-th food component and the j-th food component be c ij , then c ij The calculation formula is:

[0083] c ij =d i ·d j ;

[0084] If the deviation rate of food ingredient d j is larger, and the deviation rate d of the jth food ingredient j is also larger, and the product of the two is c ij This quantitative approach breaks the traditional pattern of analyzing components in isolation and fully considers the correlation between components.

[0085] Furthermore, the interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and combined with the weighted deviation, the comprehensive deviation rate is obtained as follows:

[0086] Assume that the comprehensive deviation rate is S and the weighted deviation of the jth food component is B j , the calculation formula of S is:

[0087]

[0088] Direct accumulation of weighted deviations for all food components;

[0089] The interaction between any two different components is expressed by the interaction coefficient c ij coupling effects that amplify or reduce deviations;

[0090] The comprehensive deviation rate generated by the comprehensive deviation rate generation module integrates the standard content of food ingredients, the actual content of food ingredients, and food ingredients. It not only includes the weighted deviation of individual food ingredients, but also determines the interaction coefficient of each ingredient in the food composition through the ingredient fuzzy deviation rate, forming a quantitative indicator covering the overall deviation degree of the food ingredient system. It can accurately capture the synergistic enhancement or antagonistic offset effects between ingredients, thereby more comprehensively assessing food quality risks, and effectively solving the defect of traditional assessment methods that ignore the mutual influence between ingredients.

[0091] The comprehensive deviation rate generation module effectively captures the ambiguity of food ingredient content through fuzzy mathematics theory, forms differentiated influences through the final weights of each ingredient in the food ingredients as differentiated weights, and forms the collaborative deviation risk between ingredients through the interaction coefficients of each ingredient in the food ingredients, thereby improving the calculation accuracy of the comprehensive deviation rate. It systematically solves the problems of insufficient uncertainty processing, vague identification of key factors, and neglect of collaborative risks in food ingredient quality assessment, and improves the comprehensiveness and accuracy of quality risk prediction in subsequent quality prediction models.

[0092] Furthermore, in the quality prediction model module, a quality risk prediction model is constructed as follows:

[0093] Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient testing data obtained by the comprehensive deviation rate generation module, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rules;

[0094] Environment-quality association mining is an existing technology. Association rule mining is a method of discovering interesting associations from large amounts of data. By mining the associations between different attributes in the data, it reveals the hidden patterns and knowledge in the data. In environment-quality association mining, the goal is to find the association rules between environmental factors and environmental quality indicators.

[0095] A quality risk prediction model is constructed using the random forest algorithm based on the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient testing data, and the comprehensive deviation rate per unit time in the future;

[0096] In the quality risk prediction model:

[0097] When the input is the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient testing data, the output of the quality risk prediction model is the current quality risk probability;

[0098] When the input is the comprehensive deviation rate per unit time in the future, production environment parameters, production equipment operation data, and food ingredient testing data, the output of the quality risk prediction model is the quality risk probability per unit time in the future.

[0099] The random forest algorithm improves the accuracy and stability of predictions by constructing multiple decision trees and aggregating the results. In this system, the random forest algorithm can effectively integrate the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient testing data to capture the complex interactions between various factors. Specifically, when the comprehensive deviation rate, real-time production environment parameters, equipment operation data, and food ingredient testing data are input, the quality risk prediction model calculates and outputs the current quality risk probability, which intuitively reflects the possibility of food quality not meeting standards under the current production status; when the comprehensive deviation rate per unit time in the future, predicted production environment parameters, equipment operation data, etc. are input, the quality risk prediction model outputs the quality risk probability per unit time in the future, helping companies to predict quality trends in advance.

[0100] In the quality prediction model module, the current quality risk probability output by the quality risk prediction model is set as P1, and the quality risk probability per unit time in the future is set as P2;

[0101] When P1 is greater than or equal to the set threshold, the alarm signal is triggered;

[0102] When P2 is greater than or equal to the set threshold, an early warning signal is triggered.

[0103] Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient detection data obtained by the comprehensive deviation rate generation module, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rules, and the random forest algorithm is used to construct a quality risk prediction model, which improves the accuracy and reliability of quality risk prediction.

[0104] like Figure 3 A food quality prediction method based on big data and fuzzy mathematics is shown, comprising the following steps:

[0105] S1. Collect multi-source data from the entire food production process;

[0106] Multi-source data includes food ingredient standard content, food ingredient actual content, food ingredients, production environment parameters, production equipment operation data, and food ingredient testing data;

[0107] S2, cleans and classifies the multi-source data collected in S1;

[0108] S3, obtaining the multi-source data after cleaning and classification in S2, constructing a fuzzy set of food ingredient standard content using fuzzy mathematics theory, defining the value range of the fuzzy set through a trapezoidal membership function, and forming a fuzzy standard content;

[0109] The difference between the actual content of food ingredients and the fuzzy standard content is quantified by the Hamming fuzzy distance formula to obtain the ingredient fuzzy deviation rate;

[0110] The importance weight of each component in food composition is expressed by triangular fuzzy number, the final weight of each component in food composition is determined by fuzzy comprehensive evaluation method, and the weighted deviation of food composition is obtained by combining the fuzzy deviation rate of the component.

[0111] The interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and the comprehensive deviation rate is obtained by combining the weighted deviation of the food components;

[0112] S4. Based on the comprehensive deviation rate obtained in S3, production environment parameters, production equipment operation data, and food ingredient testing data, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rule;

[0113] Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient detection data and the comprehensive deviation rate per unit time in the future, a quality risk prediction model is constructed using the random forest algorithm. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability per unit time in the future.

[0114] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0115] In the embodiments provided by the present invention, it should be understood that the disclosed systems or methods can be implemented in other ways. For example, the embodiments of the invention described above are merely illustrative. For example, the division of modules is only a logical function division, and other division methods may be used in actual implementation.

[0116] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected to achieve the objectives of this embodiment based on actual needs.

[0117] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or hardware plus software functional modules.

[0118] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the basic characteristics of the present invention.

[0119] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A food quality prediction system based on big data and fuzzy mathematics, characterized by: It includes: Data acquisition module: used to acquire multi-source data of the entire food production process; The multi-source data includes the standard content of food ingredients, the actual content of food ingredients, food ingredients, production environment parameters, production equipment operation data, and food ingredient detection data; Data processing module: used to clean and classify the multi-source data acquired by the data acquisition module; Comprehensive deviation rate generation module: obtains the multi-source data that has been cleaned and classified in the data processing module; Constructs a fuzzy set of the standard content of food ingredients using fuzzy mathematics theory, defines the value range of the fuzzy set through a trapezoidal membership function, and forms a fuzzy standard content; Quantifies the difference between the actual content of food ingredients and the fuzzy standard content through the Hamming fuzzy distance formula to obtain the component fuzzy deviation rate; Represents the importance weights of each component in food ingredients with triangular fuzzy numbers, determines the final weights of each component in food ingredients through the fuzzy comprehensive evaluation method, and combines the component fuzzy deviation rate to obtain the weighted deviation of food ingredients; Determines the interaction coefficient of each component in food ingredients through the component fuzzy deviation rate, and combines the weighted deviation of food ingredients to obtain the comprehensive deviation rate; Quality prediction model module: constructs a quality risk prediction model through the comprehensive deviation rate obtained by the comprehensive deviation rate generation module and the multi-source data that has been cleaned and classified in the data processing module. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability in the future unit time.

2. The food quality prediction system based on big data and fuzzy mathematics according to claim 1 is characterized in that: In the comprehensive deviation rate generation module, the value range of the fuzzy set is defined through the trapezoidal membership function as follows: where a, b, c, d are the boundary parameters of the fuzzy interval, and 0 < a < b < c < d, x is the actual content of food ingredients, and μ(x) is the membership degree.

3. The food quality prediction system based on big data and fuzzy mathematics according to claim 2, characterized in that: In the comprehensive deviation rate generation module, the weighted deviation of food ingredients is generated by combining the final weights of each component in food ingredients with the component fuzzy deviation rate in the following way: Assume that the number of food ingredients is n, i = 1, 2, 3, ..., n, w i is the final weight of the i-th food component; Assume that the fuzzy deviation rate of the i-th food ingredient is d i ; The weighted deviation B of the i-th food component i The calculation formula is: B i =w i ·d i 。 4. The food quality prediction system based on big data and fuzzy mathematics according to claim 3 is characterized in that: In the comprehensive deviation rate generation module, the interaction coefficient of each component in food ingredients is determined through the component fuzzy deviation rate as follows: Assume that the fuzzy deviation rate of the jth food ingredient is d j , j≠i; Let the interaction coefficient between the i-th food component and the j-th food component be c ij , then c ij The calculation formula is: c ij =d i ·d j。 5. The food quality prediction system based on big data and fuzzy mathematics according to claim 4, characterized in that: Determines the interaction coefficient of each component in food ingredients through the component fuzzy deviation rate, and combines the weighted deviation to obtain the comprehensive deviation rate as follows: Assume that the comprehensive deviation rate is S and the weighted deviation of the jth food component is B j , the calculation formula of S is:

6. The food quality prediction system based on big data and fuzzy mathematics according to claim 1, characterized in that: In the quality prediction model module, a quality risk prediction model is constructed as follows: According to the comprehensive deviation rate obtained by the comprehensive deviation rate generation module, production environment parameters, production equipment operation data, and food ingredient detection data, the comprehensive deviation rate in the future unit time is obtained through the environment-quality association mining rule; According to the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient detection data, and the comprehensive deviation rate in the future unit time, a quality risk prediction model is constructed through the random forest algorithm; In the quality risk prediction model: When the input is the comprehensive deviation rate, production environment parameters, production equipment operation data, and food ingredient detection data, the output of the quality risk prediction model is the current quality risk probability; When the input is the comprehensive deviation rate in the future unit time, production environment parameters, production equipment operation data, and food ingredient detection data, the output of the quality risk prediction model is the quality risk probability in the future unit time.

7. The food quality prediction system based on big data and fuzzy mathematics according to claim 6, characterized in that: In the quality prediction model module, the current quality risk probability output by the quality risk prediction model is set as P1, and the quality risk probability per unit time in the future is set as P2; When P1 is greater than or equal to the set threshold, the alarm signal is triggered; When P2 is greater than or equal to the set threshold, an early warning signal is triggered.

8. The food quality prediction method based on big data and fuzzy mathematics according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Collect multi-source data from the entire food production process; Multi-source data includes food ingredient standard content, food ingredient actual content, food ingredients, production environment parameters, production equipment operation data, and food ingredient testing data; S2, cleans and classifies the multi-source data collected in S1; S3, obtaining the multi-source data after cleaning and classification in S2, constructing a fuzzy set of food ingredient standard content using fuzzy mathematics theory, defining the value range of the fuzzy set through a trapezoidal membership function, and forming a fuzzy standard content; The difference between the actual content of food ingredients and the fuzzy standard content is quantified by the Hamming fuzzy distance formula to obtain the ingredient fuzzy deviation rate; The importance weight of each component in food composition is expressed by triangular fuzzy number, the final weight of each component in food composition is determined by fuzzy comprehensive evaluation method, and the weighted deviation of food composition is obtained by combining the fuzzy deviation rate of the component. The interaction coefficient of each component in the food composition is determined by the component fuzzy deviation rate, and the comprehensive deviation rate is obtained by combining the weighted deviation of the food components; S4. Based on the comprehensive deviation rate obtained in S3, production environment parameters, production equipment operation data, and food ingredient testing data, the comprehensive deviation rate per unit time in the future is obtained through the environment-quality association mining rule; Based on the comprehensive deviation rate, production environment parameters, production equipment operation data, food ingredient detection data and the comprehensive deviation rate per unit time in the future, a quality risk prediction model is constructed using the random forest algorithm. The quality risk prediction model is used to output the current quality risk probability and the quality risk probability per unit time in the future.

Images