Food safety monitoring method and system based on big data

Abstract

The invention discloses a food safety monitoring method and system based on big data, and the method comprises the following steps: collecting the data of each link of food, carrying out the preprocessing of the data through a big data model, building a unique food identifier, and carrying out the data association, and forming a complete food data chain; on the basis of the food data chain, risk functions influencing food safety of all links are constructed, corresponding weights are given to different risks, a risk calculation model is constructed according to the risk integral functions and the corresponding weights, and risk accumulation calculation of all the links is carried out; constructing a multi-threshold pre-warning system, dynamically calculating a food safety threshold, comparing an accumulated risk value with a pre-warning threshold, dividing risk levels, and performing intelligent pre-warning response according to the risk levels and a response mode; information of each link is visualized and monitored in real time, and feedback information is collected to continuously improve and optimize the model. According to the invention, the collected data is processed by using the big data model, so that the food safety monitoring speed and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of food, and in particular to a food safety monitoring method and system based on big data. Background Art

[0002] The modern food industry chain encompasses multiple links, including production, processing, circulation, and sales, involving numerous participants and complex logistics pathways. From farmland to table, every link in the chain can introduce food safety risks. Traditional monitoring methods struggle to fully cover and effectively manage such a complex system. More advanced technologies are needed to integrate and analyze data from the entire chain for comprehensive monitoring. Based on this, the present invention proposes a food safety monitoring method and system based on big data. Summary of the Invention

[0003] The present invention provides a food safety monitoring method based on big data, which is characterized by comprising:

[0004] S10. Collect data from all aspects of food production, circulation, and sales, and use big data models to pre-process the data, establish unique food identifiers for data association, and form a complete food data chain;

[0005] S20. Based on the food data chain, construct a four-dimensional risk integral function for each link affecting food safety, assign corresponding weights to different risks, and build a risk calculation model based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link;

[0006] S30. Build a multi-threshold early warning system, dynamically calculate food safety thresholds, compare cumulative risk values with early warning thresholds, classify risk levels, and conduct intelligent early warning responses based on risk levels and response modes;

[0007] S40. Visualize information from each link, monitor in real time, and collect feedback to continuously improve and optimize the model.

[0008] As described above, a food safety monitoring method based on big data is described, wherein the big data model pre-processes the data to supplement missing data using a predictive filling method based on a machine learning algorithm; and natural language processing technology is used to analyze feedback information and extract key information.

[0009] As described above, a food safety monitoring method based on big data is used, in which a food data chain is established by establishing a unique food identifier, and according to the food identifier, the data of the production, processing, circulation and sales links are associated to form a complete food data chain.

[0010] As described above, a food safety monitoring method based on big data, wherein the four-dimensional risk includes geographical risk, quality dimension risk, process dimension risk, and sales dimension risk.

[0011] As described above, a food safety monitoring method based on big data is used, in which the risk value of each food is calculated based on the risk factor data and corresponding weights in each link, the risk value of a certain link is calculated, and the risk values of each link are accumulated to obtain the total risk value.

[0012] As described above, a food safety monitoring method based on big data is provided, in which the calculation of the dynamic threshold is based on the safety benchmark value of historical industry data statistics, and the safety benchmark value is dynamically adjusted according to the degree of market imbalance and the impact of regulatory policies.

[0013] As described above, a food safety monitoring method based on big data is provided, in which food is divided into different risk levels according to the cumulative risk value and the warning threshold, and is divided into three levels: low risk, medium risk, and high risk, and different levels of response are performed according to the automatic warning response function.

[0014] The present invention also provides a food safety monitoring system based on big data, comprising:

[0015] Data collection and processing module: collects data from all aspects of food production, circulation, and sales, and uses big data models to pre-process the data, establishes a unique food identifier for data association, and forms a complete food data chain;

[0016] Model building module: Based on the food data chain, a four-dimensional risk integral function is constructed for each link affecting food safety, and corresponding weights are assigned to different risks. A risk calculation model is constructed based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link;

[0017] Intelligent early warning response module: Builds a multi-threshold early warning system, dynamically calculates food safety thresholds, compares cumulative risk values with early warning thresholds, categorizes risk levels, and conducts intelligent early warning responses based on risk levels and response modes;

[0018] Visualization and improvement module: visualize information in each link, monitor in real time, and collect feedback information to continuously improve and optimize the model.

[0019] The beneficial effects achieved by the present invention are as follows: the present invention improves the speed and accuracy of food safety monitoring by using a big data model to process collected data. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0021] Figure 1 This is a flow chart of a food safety monitoring method based on big data provided in Example 1 of the present application.

[0022] Figure 2 This is a schematic diagram of a food safety monitoring system based on big data provided in Example 2 of the present application. DETAILED DESCRIPTION

[0023] The following is a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] Example 1

[0025] like Figure 1 As shown, the first embodiment of the present application provides a food safety monitoring method based on big data, including:

[0026] S10. Collect data from all aspects of food production, circulation, and sales and use big data models to pre-process the data, establish a unique food identifier for data association, and form a complete food data chain.

[0027] S11. Collect data on all aspects of food production, circulation and sales.

[0028] Collect raw material information during the production process to assess the safety of the raw materials themselves and the impact of the growing environment on the quality of the raw materials. Establish a data interface with suppliers to obtain data such as the origin, variety, pesticide residues, veterinary drug residues, and heavy metal content of the raw materials. For agricultural raw materials, use satellite remote sensing, drone monitoring, and ground sensor networks to collect environmental data such as soil fertility, irrigation water quality, and meteorological conditions. At the same time, require suppliers to regularly provide third-party inspection reports for raw materials. Collect production process parameters to understand whether the production process meets standard specifications. Install various sensors in the food production workshop, including temperature sensors, humidity sensors, pressure sensors, flow sensors, etc., to collect production parameters at each key node on the production line in real time and transmit these sensor data to the data center in real time.

[0029] Collect information on additive usage during processing. Through the company's production management system, detailed records are kept of each additive's type, dosage, addition time, and operator information. Furthermore, intelligent monitoring equipment is installed in additive storage areas to monitor additive inventory in real time and ensure that additive usage remains within safe limits. Electronic tagging technology is used to track each batch of additives, recording information such as their source and shelf life. Processing environment monitoring data is collected, and air quality monitoring equipment is installed in the processing workshop to detect airborne microbial content, dust particle counts, and hazardous gas concentrations. Microbial sampling and testing is regularly conducted on workshop floors and equipment surfaces, with laboratory analysis used to obtain data on microbial species and counts.

[0030] Collect data on the distribution process, including transportation and storage conditions. Install positioning equipment, temperature sensors, humidity sensors, and vibration sensors on transport vehicles to collect transportation condition data. Use positioning to track the vehicle's route, speed, and dwell time in real time. Temperature and humidity sensors monitor changes in temperature and humidity within the vehicle during transportation, and vibration sensors record vibrations during transportation. Install temperature and humidity control equipment and monitoring systems in the warehouse to record temperature and humidity data in different areas of the warehouse in real time. At the same time, use video surveillance systems and inventory management software to record information such as food entry and exit times, inventory quantities, and stacking locations. Through the intelligent shelving system, the storage status of food at each shelf is monitored in real time.

[0031] Collect data from the sales process, including market inspection data and consumer feedback. Local food regulatory authorities send food samples from market inspections to specialized laboratories for testing, covering physical, chemical, and microbiological indicators. Upon completion, detailed test reports are uploaded to the data center via a dedicated data interface. Build a multi-channel consumer feedback platform, including official websites, social media platforms, and complaint hotlines. Consumers can provide feedback on food quality issues through text descriptions, images, and videos. Natural language processing technology is used to analyze these textual feedback and extract key information.

[0032] S12. Use big data models to preprocess the collected data and establish unique food identification.

[0033] Collected data is thoroughly checked to remove duplicates and ensure accuracy and completeness. Erroneous data is filtered by setting reasonable thresholds, and outliers are corrected or deleted. Missing data is addressed using methods such as mean filling, median filling, and predictive filling based on machine learning algorithms.

[0034] Data from different sources and formats is converted to a standardized format to make different types of data comparable. A unified format is used for date formats. For numerical data, normalization or standardization methods are used to map the data to a specific range based on its characteristics and distribution. Categorical data, including food categories and additive types, is encoded using one-hot encoding to enable efficient processing by computer models.

[0035] Establish unique food identification, including product batch numbers and traceability codes, to link data from production, processing, distribution, and sales. Match raw material information and process parameters from the production phase with additive usage and processing environment data from the processing phase using the product batch number. In the distribution phase, use traceability codes to link transportation and storage conditions with production and processing data. In the sales phase, match market sampling data and consumer feedback with data from previous stages to form a complete food data chain.

[0036] S20. Based on the food data chain, construct a four-dimensional risk integral function for each link affecting food safety, assign corresponding weights to different risks, and build a risk calculation model based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link.

[0037] S21. Construct a four-dimensional risk integral function for each link that affects food safety.

[0038] Based on professional knowledge and past research results in the field of food safety, the key risk factors affecting food safety at each stage are determined and used to construct a four-dimensional risk integral function. The specific formula is: [t0, t1] is the time window covering the entire cycle from production to sales, the integral domain V covers the spatial range of the supply chain, and G(x, y) is the two-dimensional geographical risk field. n represents the number of raw material production areas, w i is the origin risk weight based on historical detection data, α is the distance attenuation coefficient, β is the time attenuation coefficient, d i represents the Euclidean distance between the target point (x, y) and the i-th raw material production place, t i is the logistics time delay factor, e is a natural constant; Q(M,P) is the quality dimension risk, g1 is the microbial risk steepness coefficient, M(t) is the total number of time-varying colonies, M th is the microbial exceeding threshold, m is the number of physical and chemical indicators, P i is the test value of the ith physical and chemical index, which includes pesticide residues, veterinary drug residues, additive dosage, etc. i0 is the national standard limit value of item i, is the hazard weight coefficient, p is the over-limit nonlinear index; O c(t) is the process dimension risk, λ a is the cumulative coefficient of production risk, t a is the production cycle, O c Score the operational standardization, F t is the personnel fatigue factor, D f is the equipment failure rate, ΔT is the transportation temperature control deviation, ΔT max is the maximum allowable temperature control deviation, v is the transportation vibration acceleration, and v crit is the critical acceleration of package damage, b is the nonlinear index of physical shock, χ is the logistics time risk coefficient, t l is the logistics overtime; R sales For sales dimension risk, φ is the storage risk accumulation coefficient, T c is the actual storage days, N is the number of unqualified items in random inspection, N0 is the industry average unqualified item benchmark value, ε is the nonlinear index of random inspection risk, is the complaint sensitivity coefficient, and C is the frequency of consumer complaints.

[0039] S22. Assign weights to risk factors.

[0040] By formula Evaluate the importance of each risk factor and assign corresponding weights to different risk factors. i is the final weight of the i-th risk, θ is the subjective and objective weight adjustment coefficient, obtained through the model self-learning parameters, AHP i is the weight of the hierarchical analysis method, obtained according to the eigenvector of the expert scoring matrix, E i is the information entropy value, τ is the structure-data adjustment coefficient, which is set according to the characteristics of the food industry, DF i The data credibility factor is set. The weight setting will be regularly reassessed and adjusted based on industry development, new food safety issues, etc.

[0041] S23. Construct a risk calculation model based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link.

[0042] For each food identified by a unique identifier, the risk value of each link is calculated based on its risk factor data and corresponding weights at each link, and the cumulative risk value of each link is calculated.

[0043] The specific formula is Where K is the total number of links, is the total risk of food individual j, W k,i is the weight of the i-th risk factor in the k-th link, is the i-th risk factor value of the k-th link, ξ k is the basic coupling coefficient, is the risk gradient of the adjacent link, and the ReLU function ensures that only the positive risk gradient has an impact, M k is the number of human error categories in the kth link, δ k,m is the influence coefficient of the mth type of human error in the kth link, is the time-varying human error function.

[0044] S30. Build a multi-threshold early warning system, dynamically calculate food safety thresholds, compare cumulative risk values with early warning thresholds, divide risk levels, and make intelligent early warning decisions based on risk levels and response modes.

[0045] Build a multi-threshold early warning system to promptly detect food safety hazards and take effective measures to prevent problematic foods from entering the market or further spreading. The dynamic threshold calculation formula is:

[0046] T b The safety benchmark value of the industry's historical data statistics is the 90th percentile of the risk value of the same category of food in the past three years, ψ·Φ -1 (P) is the statistical fluctuation correction term, ψ is the standard deviation of historical data, Φ -1 (P Φ ) is the inverse function of the standard normal distribution, P Φ is the confidence level, is the market sensitivity coefficient, is the absolute value of the difference between supply and demand, reflecting the degree of market imbalance, Demand(t) is the market demand function at time t, Supply(t) is the dynamic supply chain supply at time t, and Supply Avg is the average supply, η is the time decay coefficient. I is the number of regulatory policy items, is the impact coefficient of the i-th regulatory policy, Ψ is the Dirac function, at time t i 1 at t and 0 at other times, indicating that the policy i The impact is effective at all times.

[0047] According to the cumulative risk value R total and warning threshold T threshold Based on the size of the risk, foods are classified into three risk levels: low risk, medium risk, and high risk. When setting the risk level threshold, reference is made to the degree of correlation between foods of different risk levels and actual food safety issues in historical data to ensure that the risk level classification has practical guiding significance.

[0048] When a food product's risk level reaches medium or high, the system automatically triggers a response alert. For medium-risk foods, the system notifies the food manufacturer and regulatory authorities via text messages, emails, and other means, reminding the manufacturer to conduct a self-inspection of the production process, and for regulatory authorities to increase the frequency of random inspections of that manufacturer or batch of food. For high-risk foods, in addition to the aforementioned notifications, a red alert is issued on the monitoring platform, requiring the manufacturer to immediately halt production and sales, initiate a recall, and for regulatory authorities to conduct a comprehensive investigation. The alert includes detailed information about the food, the reason for the increased risk level, and recommended measures.

[0049] The automatic warning response function is Among them, A is the response amplitude, which is set according to the risk level, ι is the response rate coefficient, which is used to control the speed at which early warning measures take effect, and t d is the current time, starting from the occurrence of the risk event, t d0 The response inflection point time is the critical point from the triggering of the warning to the effectiveness of the measures. In response to the pattern matching function, Mode m To match the mth response pattern in the response pattern, B is the baseline response intensity, maintaining the minimum response level for daily monitoring.

[0050] S40. Visualize information from each link, monitor in real time, and collect feedback to continuously improve and optimize the model.

[0051] Newly collected and processed data is fed into the risk calculation model in real time, dynamically updating the cumulative risk value and risk level of food products. A data dashboard is established to visualize the real-time risk status of different foods, including information such as food name, batch, risk value at each stage, cumulative risk value, and risk level. Monitoring personnel can access food risk status anytime and anywhere through computers, mobile phones, and other terminals, keeping abreast of changes in food risk at every stage and supporting rapid response.

[0052] Collect feedback data from the implementation of risk monitoring and early warning measures, including company self-inspection results, regulatory inspection reports, and test data on recalled foods. This feedback data will be incorporated back into the data collection process to update and supplement the existing data. If new risk factors are discovered during the testing of recalled foods, they will be added to the risk factor database and the risk values of the relevant foods will be recalculated. This will enable the risk assessment model to continuously adapt to changes in actual conditions and improve the accuracy of risk assessments.

[0053] Regularly evaluate the risk value calculation model, using a variety of evaluation metrics, including accuracy and recall, to analyze the model's performance in risk grading and early warning. Based on the evaluation results, adjust and optimize the selection of risk factors, the setting of weights, and the calculation method. If a risk factor is found to be insignificant in the actual risk assessment, its weight will be reduced or considered for removal. If a new risk factor appears frequently and has a significant impact on food safety, it will be included in the risk factors and assigned an appropriate weight.

[0054] Example 2

[0055] like Figure 2 As shown, the second embodiment of the present application provides a food safety monitoring system based on big data, including:

[0056] Data collection and processing module: includes collection submodule and processing submodule.

[0057] Collection submodule: used to collect data from all aspects of food production, circulation and sales.

[0058] Collect raw material information during the production process to assess the safety of the raw materials themselves and the impact of the growing environment on the quality of the raw materials. Establish a data interface with suppliers to obtain data such as the origin, variety, pesticide residues, veterinary drug residues, and heavy metal content of the raw materials. For agricultural raw materials, use satellite remote sensing, drone monitoring, and ground sensor networks to collect environmental data such as soil fertility, irrigation water quality, and meteorological conditions. At the same time, require suppliers to regularly provide third-party inspection reports for raw materials. Collect production process parameters to understand whether the production process meets standard specifications. Install various sensors in the food production workshop, including temperature sensors, humidity sensors, pressure sensors, flow sensors, etc., to collect production parameters at each key node on the production line in real time and transmit these sensor data to the data center in real time.

[0059] Collect information on additive usage during processing. Through the company's production management system, detailed records are kept of each additive's type, dosage, addition time, and operator information. Furthermore, intelligent monitoring equipment is installed in additive storage areas to monitor additive inventory in real time and ensure that additive usage remains within safe limits. Electronic tagging technology is used to track each batch of additives, recording information such as their source and shelf life. Processing environment monitoring data is collected, and air quality monitoring equipment is installed in the processing workshop to detect airborne microbial content, dust particle counts, and hazardous gas concentrations. Microbial sampling and testing is regularly conducted on workshop floors and equipment surfaces, with laboratory analysis used to obtain data on microbial species and counts.

[0060] Collect data on the distribution process, including transportation and storage conditions. Install positioning equipment, temperature sensors, humidity sensors, and vibration sensors on transport vehicles to collect transportation condition data. Use positioning to track the vehicle's route, speed, and dwell time in real time. Temperature and humidity sensors monitor changes in temperature and humidity within the vehicle during transportation, and vibration sensors record vibrations during transportation. Install temperature and humidity control equipment and monitoring systems in the warehouse to record temperature and humidity data in different areas of the warehouse in real time. At the same time, use video surveillance systems and inventory management software to record information such as food entry and exit times, inventory quantities, and stacking locations. Through the intelligent shelving system, the storage status of food at each shelf is monitored in real time.

[0061] Collect data from the sales process, including market inspection data and consumer feedback. Local food regulatory authorities send food samples from market inspections to specialized laboratories for testing, covering physical, chemical, and microbiological indicators. Upon completion, detailed test reports are uploaded to the data center via a dedicated data interface. Build a multi-channel consumer feedback platform, including official websites, social media platforms, and complaint hotlines. Consumers can provide feedback on food quality issues through text descriptions, images, and videos. Natural language processing technology is used to analyze these textual feedback and extract key information.

[0062] Processing submodule: Use big data models to pre-process the collected data and establish a unique food identifier.

[0063] Collected data is thoroughly checked to remove duplicates and ensure accuracy and completeness. Erroneous data is filtered by setting reasonable thresholds, and outliers are corrected or deleted. Missing data is addressed using methods such as mean filling, median filling, and predictive filling based on machine learning algorithms.

[0064] Data from different sources and formats is converted to a standardized format to make different types of data comparable. A unified format is used for date formats. For numerical data, normalization or standardization methods are used to map the data to a specific range based on its characteristics and distribution. Categorical data, including food categories and additive types, is encoded using one-hot encoding to enable efficient processing by computer models.

[0065] Establish unique food identification, including product batch numbers and traceability codes, to link data from production, processing, distribution, and sales. Match raw material information and process parameters from the production phase with additive usage and processing environment data from the processing phase using the product batch number. In the distribution phase, use traceability codes to link transportation and storage conditions with production and processing data. In the sales phase, match market sampling data and consumer feedback with data from previous stages to form a complete food data chain.

[0066] Build model modules: including risk sub-module, weight sub-module, and model sub-module.

[0067] Risk sub-module: used to construct a four-dimensional risk integral function for each link affecting food safety.

[0068] Based on professional knowledge and past research results in the field of food safety, the key risk factors affecting food safety at each stage are determined and used to construct a four-dimensional risk integral function. The specific formula is: [t0, t1] is the time window covering the entire cycle from production to sales, the integral domain V covers the spatial range of the supply chain, and G(x, y) is the two-dimensional geographical risk field. n represents the number of raw material production areas, w i is the origin risk weight based on historical detection data, α is the distance attenuation coefficient, β is the time attenuation coefficient, d i represents the Euclidean distance between the target point (x, y) and the i-th raw material production place, t i is the logistics time delay factor, e is a natural constant; Q(M,P) is the quality dimension risk, g1 is the microbial risk steepness coefficient, M(t) is the total number of time-varying colonies, M th is the microbial exceeding threshold, m is the number of physical and chemical indicators, P i is the test value of the ith physical and chemical index, which includes pesticide residues, veterinary drug residues, additive dosage, etc. i0 is the national standard limit value of item i, is the hazard weight coefficient, p is the over-limit nonlinear index; O c (t) is the process dimension risk, λ a is the cumulative coefficient of production risk, t a is the production cycle, O c Score the operational standardization, F t is the personnel fatigue factor, D f is the equipment failure rate, ΔT is the transportation temperature control deviation, ΔT max is the maximum allowable temperature control deviation, v is the transportation vibration acceleration, and v critis the critical acceleration of package damage, b is the nonlinear index of physical shock, χ is the logistics time risk coefficient, t l is the logistics overtime; R sales For sales dimension risk, φ is the storage risk accumulation coefficient, T c is the actual storage days, N is the number of unqualified items in random inspection, N0 is the industry average unqualified item benchmark value, ε is the nonlinear index of random inspection risk, is the complaint sensitivity coefficient, and C is the frequency of consumer complaints.

[0069] Weight submodule: used to assign weights to risk factors.

[0070] By formula Evaluate the importance of each risk factor and assign corresponding weights to different risk factors. i is the final weight of the i-th risk, θ is the subjective and objective weight adjustment coefficient, obtained through the model self-learning parameters, AHP i is the weight of the hierarchical analysis method, obtained according to the eigenvector of the expert scoring matrix, E i is the information entropy value, τ is the structure-data adjustment coefficient, which is set according to the characteristics of the food industry, DF i The data credibility factor is set. The weight setting will be regularly reassessed and adjusted based on industry development, new food safety issues, etc.

[0071] Model submodule: Construct a risk calculation model based on the risk integral function and corresponding weights, and perform cumulative risk calculations for each link.

[0072] For each food identified by a unique identifier, the risk value of each link is calculated based on its risk factor data and corresponding weights at each link, and the cumulative risk value of each link is calculated.

[0073] The specific formula is Where K is the total number of links, is the total risk of food individual j, W k,i is the weight of the i-th risk factor in the k-th link, is the i-th risk factor value of the k-th link, ξ k is the basic coupling coefficient, is the risk gradient of the adjacent link, and the ReLU function ensures that only the positive risk gradient has an impact, M k is the number of human error categories in the kth link, δ k,m is the influence coefficient of the mth type of human error in the kth link, is the time-varying human error function.

[0074] Intelligent early warning response module: Builds a multi-threshold early warning system, dynamically calculates food safety thresholds, compares cumulative risk values with early warning thresholds, categorizes risk levels, and conducts intelligent early warning responses based on risk levels and response modes;

[0075] Build a multi-threshold early warning system to promptly detect food safety hazards and take effective measures to prevent problematic foods from entering the market or further spreading. The dynamic threshold calculation formula is:

[0076] T b The safety benchmark value of the industry's historical data statistics is the 90th percentile of the risk value of the same category of food in the past three years, ψ·Φ -1 (P) is the statistical fluctuation correction term, ψ is the standard deviation of historical data, Φ -1 (P Φ ) is the inverse function of the standard normal distribution, P Φ is the confidence level, is the market sensitivity coefficient, is the absolute value of the difference between supply and demand, reflecting the degree of market imbalance, Demand(t) is the market demand function at time t, Supply(t) is the dynamic supply chain supply at time t, and Supply Avg is the average supply, η is the time decay coefficient. I is the number of regulatory policy items, is the impact coefficient of the i-th regulatory policy, Ψ is the Dirac function, at time t i 1 at t and 0 at other times, indicating that the policy i The impact is effective at all times.

[0077] According to the cumulative risk value R total and warning threshold T threshold Based on the size of the risk, foods are classified into three risk levels: low risk, medium risk, and high risk. When setting the risk level threshold, reference is made to the degree of correlation between foods of different risk levels and actual food safety issues in historical data to ensure that the risk level classification has practical guiding significance.

[0078] When a food product's risk level reaches medium or high, the system automatically triggers a response alert. For medium-risk foods, the system notifies the food manufacturer and regulatory authorities via text messages, emails, and other means, reminding the manufacturer to conduct a self-inspection of the production process, and for regulatory authorities to increase the frequency of random inspections of that manufacturer or batch of food. For high-risk foods, in addition to the aforementioned notifications, a red alert is issued on the monitoring platform, requiring the manufacturer to immediately halt production and sales, initiate a recall, and for regulatory authorities to conduct a comprehensive investigation. The alert includes detailed information about the food, the reason for the increased risk level, and recommended measures.

[0079] The automatic warning response function is Among them, A is the response amplitude, which is set according to the risk level, ι is the response rate coefficient, which is used to control the speed at which early warning measures take effect, and t d is the current time, starting from the occurrence of the risk event, t d0 The response inflection point time is the critical point from the triggering of the warning to the effectiveness of the measures. In response to the pattern matching function, Mode m To match the mth response pattern in the response pattern, B is the baseline response intensity, maintaining the minimum response level for daily monitoring.

[0080] Visualization and improvement module: visualize information in each link, monitor in real time, and collect feedback information to continuously improve and optimize the model.

[0081] Newly collected and processed data is fed into the risk calculation model in real time, dynamically updating the cumulative risk value and risk level of food products. A data dashboard is established to visualize the real-time risk status of different foods, including information such as food name, batch, risk value at each stage, cumulative risk value, and risk level. Monitoring personnel can access food risk status anytime and anywhere through computers, mobile phones, and other terminals, keeping abreast of changes in food risk at every stage and supporting rapid response.

[0082] Collect feedback data from the implementation of risk monitoring and early warning measures, including company self-inspection results, regulatory inspection reports, and test data on recalled foods. This feedback data will be incorporated back into the data collection process to update and supplement the existing data. If new risk factors are discovered during the testing of recalled foods, they will be added to the risk factor database and the risk values of the relevant foods will be recalculated. This will enable the risk assessment model to continuously adapt to changes in actual conditions and improve the accuracy of risk assessments.

[0083] Regularly evaluate the risk value calculation model, using a variety of evaluation metrics, including accuracy and recall, to analyze the model's performance in risk grading and early warning. Based on the evaluation results, adjust and optimize the selection of risk factors, the setting of weights, and the calculation method. If a risk factor is found to be insignificant in the actual risk assessment, its weight will be reduced or considered for removal. If a new risk factor appears frequently and has a significant impact on food safety, it will be included in the risk factors and assigned an appropriate weight.

[0084] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A food safety monitoring method and system based on big data, characterized in that: include: S10. Collect data from all aspects of food production, circulation, and sales, and use big data models to pre-process the data, establish unique food identifiers for data association, and form a complete food data chain; S20. Based on the food data chain, construct a four-dimensional risk integral function for each link affecting food safety, assign corresponding weights to different risks, and build a risk calculation model based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link; S30. Build a multi-threshold early warning system, dynamically calculate food safety thresholds, compare cumulative risk values with early warning thresholds, classify risk levels, and conduct intelligent early warning responses based on risk levels and response modes; S40. Visualize information from each link, monitor in real time, and collect feedback to continuously improve and optimize the model.

2. A food safety monitoring method based on big data according to claim 1, characterized in that: The big data model pre-processes the data to supplement missing data using a predictive filling method based on a machine learning algorithm; and uses natural language processing technology to analyze feedback information and extract key information.

3. A food safety monitoring method based on big data according to claim 1, characterized in that: The food data chain is established by establishing a unique food identification, and linking the data of production, processing, circulation and sales links according to the food identification to form a complete food data chain.

4. A food safety monitoring method based on big data according to claim 1, characterized in that: The four-dimensional risks include geographical risks, quality-dimension risks, process-dimension risks, and sales-dimension risks.

5. The food safety monitoring method based on big data according to claim 1, characterized in that: The risk value of each food is calculated based on the risk factor data and corresponding weights in each link, and the risk values of each link are accumulated to obtain the total risk value.

6. A food safety monitoring method based on big data according to claim 1, characterized in that: The calculation of the dynamic threshold is based on the safety benchmark value of historical industry data statistics, and the safety benchmark value is dynamically adjusted according to the degree of market imbalance and the impact of regulatory policies.

7. The food safety monitoring method based on big data according to claim 1, characterized in that: According to the cumulative risk value and warning threshold, food is divided into different risk levels, namely low risk, medium risk and high risk, and different levels of response are carried out according to the automatic warning response function.

8. A food safety monitoring system based on big data, characterized in that: include: Data collection and processing module: collects data from all aspects of food production, circulation, and sales, and uses big data models to pre-process the data, establishes a unique food identifier for data association, and forms a complete food data chain; Model building module: Based on the food data chain, a four-dimensional risk integral function is constructed for each link affecting food safety, and corresponding weights are assigned to different risks. A risk calculation model is constructed based on the risk integral function and corresponding weights to perform cumulative risk calculations for each link; Intelligent early warning response module: Builds a multi-threshold early warning system, dynamically calculates food safety thresholds, compares cumulative risk values with early warning thresholds, categorizes risk levels, and conducts intelligent early warning responses based on risk levels and response modes; Visualization and improvement module: visualize information in each link, monitor in real time, and collect feedback information to continuously improve and optimize the model.

Images