A food safety traceability and risk control system based on data mining
The food safety traceability and risk control system based on data mining solves the problems of omissions in traceability scope and unscientific selection of risk nodes in traditional systems in complex supply chains, and achieves accurate traceability and efficient management, ensuring the integrity and accuracy of food safety traceability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGXI LINCHUANG CATERING CULTURE CO LTD
- Filing Date
- 2026-02-09
- Publication Date
- 2026-05-29
AI Technical Summary
Traditional food safety traceability systems are ill-suited to complex supply chain scenarios and cannot fully cover batch splitting and multi-node cross-linking, resulting in omissions in traceability scope, inaccurate risk tracing, and a lack of scientific rigor in risk node selection, leading to ineffective control measures or wasted resources.
The food safety traceability and risk control system based on data mining initializes the supply chain node library, forms a node-edge topology network, quantifies the business relationships and risk transmission potential between nodes, accurately locks the risk source links, generates a priority-ranked core traceability sequence, and achieves precise traceability and targeted control.
To ensure that no aspect of the traceability is missed, to accurately pinpoint the source of risk, to achieve efficient and precise risk control, to avoid wasting resources, and to meet the needs of rapid location and accurate traceability.
Smart Images
Figure CN122114950A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent food safety management technology, and more specifically, to a food safety traceability and risk control system based on data mining. Background Technology
[0002] Currently, traditional food safety traceability systems are ill-suited to the complex flow scenarios of the food supply chain. Most systems lack systematic modeling of the relationships between supply chain nodes, fail to effectively integrate key information such as node business attributes and flow sequence, and cannot fully cover the actual business needs of batch splitting and multi-node cross-link flow. This can easily lead to omissions in the traceability scope and make it difficult to completely reconstruct the entire flow trajectory of food from source to end, thereby affecting the completeness and accuracy of risk traceability.
[0003] At the level of risk link identification, existing technologies mostly rely on qualitative judgments or single-dimensional data assessments of link risks, lacking comprehensive quantitative analysis of the impact of risk transmission synergy and flow timeliness. Traditional methods struggle to accurately distinguish the risk levels of different traceability links and cannot quickly identify core links with concentrated risks. This often leads to generalized investigations during the traceability process, which is not only inefficient but may also delay risk control opportunities due to a lack of focus, failing to meet the core needs of rapid location and accurate traceability in food safety traceability.
[0004] The identification and prioritization of core risk nodes is another weakness of existing traceability technologies. Traditional systems often focus on investigating nodes within a single link, neglecting the transmission and diffusion of risks originating from the same source across links. They fail to consider key factors such as the connection strength of nodes within the entire supply chain network and the frequency of cross-link sharing, resulting in a lack of scientific rigor in risk node selection. This leads to a lack of clear targets for subsequent risk management measures, either resulting in indiscriminate control and wasted resources, or ineffective intervention at high-risk nodes due to confused management priorities, thus affecting the accuracy and efficiency of risk control. Summary of the Invention
[0005] In view of the shortcomings of existing technologies, the purpose of this invention is to provide a food safety traceability and risk control system based on data mining.
[0006] To achieve the above objectives, the present invention provides the following technical solution:
[0007] A food safety traceability and risk control system based on data mining includes a supply chain node library initialization module, a risk link marking module, and a food safety traceability risk control module;
[0008] The supply chain node library initialization module is used to initialize the supply chain node library and label the geographical coordinates, business attributes, and detection indicator sets of each node in the supply chain node library.
[0009] The risk link marking module, upon receiving a food safety traceability task, locks down the range of nodes associated with the traceability task, designates all nodes within this range as network nodes, and identifies direct business relationships between nodes as network edges, forming a node-edge topology network. It then obtains the dynamic correlation degree of each associated edge in the topology network. When the dynamic correlation degree is higher than the dynamic correlation threshold and the nodes on both sides of the correlation edge are upstream and downstream business attributes, the correlation edge is marked as a core correlation edge. The batch time sequence of the core correlation edge is obtained. The core correlation edges are connected according to the batch time sequence to form multiple food safety traceability links. The high-probability initial screening index of each food safety traceability link is obtained. The food safety traceability link with the highest high-probability initial screening index is marked as the risk source link, and the rest of the food safety traceability links are marked as non-source links.
[0010] The food safety traceability and risk control module generates a core food safety traceability sequence based on the risk source link and non-source link, and then performs traceability and graded circuit breaker measures on high-risk candidate nodes in sequence according to the core food safety traceability sequence.
[0011] Furthermore, after receiving a food safety traceability task, the scope of nodes associated with the traceability task is locked in the following steps: After receiving a food safety traceability task, the task tags are matched through the rule engine, and the preset task-node attribute mapping table is automatically associated. Through the food traceability code associated with the task, the full-chain node IDs bound to the traceability code in the supply chain node library are retrieved, and the target set of nodes for which data needs to be collected is locked in, which is the scope of nodes associated with the traceability task.
[0012] Furthermore, the dynamic correlation between the two nodes The acquisition process is as follows: Label the two nodes of the associated edge as node i and node j respectively, and obtain the normalized geographical distance between node i and node j. Obtain the normalized batch time difference between node i and node j. Obtain the degree centrality of node i. Get node j Through formula ;in, , All are weighting coefficients. .
[0013] Furthermore, the high-probability initial screening index for risks in the food safety traceability chain is obtained through the following steps: Select a food safety traceability chain, obtain the risk transmission consistency index for every two adjacent nodes in the chain, and sum and average all risk transmission consistency indices to calculate the average risk transmission consistency index. To obtain the timeliness deviation between every two adjacent nodes in the food safety traceability chain. The average timeliness deviation is calculated by summing and averaging all the timeliness deviations. Through formula The high-probability initial screening index for food safety traceability links was calculated. ;in, , All are weighting coefficients. .
[0014] Furthermore, the timeliness deviation between any two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them node a and node b respectively, and obtain the actual processing time of node a and node b. Obtain the business attributes of node a and node b, and determine the standard transfer time based on the business attributes between node a and node b. Through formula The time difference between two adjacent nodes was calculated. .
[0015] Furthermore, the risk transmission consistency index between every two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them as node a and node b respectively; obtain the detection index sets for nodes a and b; filter out the detection data in the detection index set of node a that are associated with the food safety traceability task, and label them as the filtered index set. The detection data associated with the food safety traceability task in the detection index set of node b are selected and marked as the selected index set. The business attributes of node a and node b are obtained. Based on the business attributes between node a and node b, multiple related indicator pairs are determined from two sets of screening indicators. The abnormal coordination consistency index Pac of each related indicator pair is obtained. The average of the abnormal coordination consistency indices of all related indicator pairs is summed to calculate the risk transmission consistency index of two adjacent nodes.
[0016] Furthermore, the abnormal consistency index Pac of the associated indicator pair is obtained through the following steps: Select an associated indicator pair, obtain the safe upper threshold and safe lower threshold for the two detection data in the associated indicator pair, obtain the abnormal performance index API for each detection data, sum the abnormal performance indices of the two detection data to calculate the abnormal performance superposition index PCA, calculate the absolute difference between the abnormal performance indices of the two detection data to calculate the abnormal performance difference index APD, and then use the formula... The abnormal synergy index Pac of the associated index pair is calculated.
[0017] Furthermore, the core traceability nodes for food safety are determined based on the risk source links and non-source links. The steps are as follows: select risk source links, screen out high-risk candidate nodes in the risk source links, determine the core traceability index of each high-risk candidate node based on the non-source links, sort all high-risk candidate nodes in descending order of the core traceability index value, and generate a core traceability sequence for food safety based on the sorting.
[0018] Furthermore, high-risk candidate nodes in the risk source links are screened out. The steps are as follows: obtain the source link contribution value of each node in the risk source links. When the source link contribution value is higher than the source link contribution threshold, the node is marked as a high-risk candidate node. Then, all high-risk candidate nodes are counted.
[0019] The source link contribution value of a node is obtained through the following steps: Select a node, obtain the set of screening indicators for that node, and then obtain the anomaly performance index for each detection data in the set of screening indicators. Sum and average the anomaly performance indices of all detection data to calculate the anomaly performance contribution value Vfpc. Obtain the two risk transmission consistency indices of the two adjacent nodes, sum and average the two risk transmission consistency indices to calculate the adjacent anomaly contribution value ksd. Then, use the formula... The source link contribution value of this node is calculated. ;in, , All are weighting coefficients. .
[0020] Furthermore, the core traceability index of high-risk candidate nodes is determined through the following steps: Select a high-risk candidate node and obtain its degree centrality in the network topology. Each time a high-risk candidate node appears in a non-source link, the non-source link sharing frequency is incremented by one. Finally, the non-source link sharing frequencies are summed and marked as... The total number of non-source links is marked as M, and then calculated using the formula... The core traceability index of high-risk candidate nodes was calculated. ;in, , All are weighting coefficients. .
[0021] Compared with the prior art, the present invention has the following beneficial effects:
[0022] The system of this invention forms a complete node-edge topology network by treating traceability task-related nodes as network nodes and direct business relationships between nodes as network edges. It then uses dynamic correlation to accurately quantify the business relationships and risk transmission potential between nodes, comprehensively covering complex scenarios in the food supply chain involving batch splitting and multi-node flow, ensuring no omissions in traceability. Through a high-probability initial risk screening index, it achieves a quantitative assessment of the risk level of each traceability link, objectively reflecting the synergy of risk transmission in the link and the catalytic effect of flow timeliness on risk. This allows for precise identification of risk source links and non-source links, quickly focusing on the core links with the highest risk level, providing a clear direction for subsequent risk tracing.
[0023] By screening high-risk candidate nodes in risk source links using source link contribution values, the core risk attributes of candidate nodes are ensured. Then, by combining the sharing frequency of non-source links and the network connection strength of nodes, a core traceability index is calculated, ultimately forming a priority-ranked core traceability sequence. This takes into account the impact of cross-link same-source risks, extending the traceability focus from a single node to key nodes in multiple links. This achieves a scientific ranking of risk node priorities, providing precise targeting for subsequent risk control, allowing control measures to be implemented in an orderly manner according to priority, avoiding resource waste, and achieving efficient and precise risk control. Attached Figure Description
[0024] Fig. 1 This is a schematic diagram of a food safety traceability and risk control system based on data mining.
[0025] Fig. 2 A flowchart for obtaining the risk propagation consistency index between two adjacent nodes;
[0026] Fig. 3 This is a flowchart for obtaining the abnormal coordination consistency index of related indicator pairs. Detailed Implementation
[0027] Reference Figs. 1 to 3 A food safety traceability and risk control system based on data mining includes a supply chain node library initialization module, a risk link marking module, and a food safety traceability risk control module.
[0028] The supply chain node library initialization module initializes the supply chain node library (independent entities in the entire food supply chain, such as planting base A, processing plant B, and cold chain warehouse C). It marks the geographical coordinates, business attributes (such as planting attributes, processing attributes, and storage attributes) and detection index sets of each node in the supply chain node library. (Nodes with different business attributes correspond to different detection index sets. For example, the detection index set of planting attribute nodes includes input detection data (pesticide residues) and environmental detection data (planting soil pH value, irrigation water heavy metal content, and field temperature and humidity); the detection index set of processing attribute nodes includes processing technology detection data (sterilization temperature and cooling rate), finished product inspection data (physicochemical indicators), and microbiological indicators (total bacterial count).
[0029] The risk link marking module, upon receiving a food safety traceability task (such as a cold chain pork spoilage risk traceability task), locks the scope of nodes associated with the traceability task. All nodes within this scope are designated as network nodes, and direct business relationships between nodes (such as pig transportation from breeding to slaughter, and cold chain distribution from slaughter to warehousing) are designated as network edges, forming a node-edge topology network. The module then obtains the dynamic correlation degree of each associated edge in the topology network. When the dynamic correlation degree is higher than the dynamic correlation threshold and the nodes on both sides of the correlation edge are upstream and downstream business attributes (the dynamic correlation threshold is set based on historical statistical data, and the correlation edges are passed through horizontal business attributes), the correlation edge is marked as a core correlation edge. The batch time sequence of the core correlation edge is obtained (e.g., PL-003 slaughter time (2025-11-01, 08:00) → P-012 slaughter time (2025-11-02, 14:00)). The core correlation edges are connected according to the batch time sequence to form multiple food safety traceability links (e.g., P-012 after slaughter). The pork was divided into two batches and stored in two frozen warehouses, W-008 and W-009, respectively. Then, corresponding to different supermarkets (S-025 and S-026), two core links were formed: PL-003→P-012→W-008→S-025 and PL-003→P-012→W-009→S-026. The high-probability initial screening index of each food safety traceability link was obtained. The food safety traceability link with the highest high-probability initial screening index was marked as the risk source link, and the remaining food safety traceability links (non-risk source links) were marked as non-source links.
[0030] After receiving a food safety traceability task, the scope of nodes associated with the traceability task is determined as follows: After receiving the food safety traceability task, the task tags (such as cold chain, pork, spoilage) are matched through the rule engine, and the preset task-node attribute mapping table is automatically associated (such as cold chain pork → planting and breeding (live pigs), production and processing (slaughter), warehousing and logistics (frozen storage), sales terminal (supermarket)). Through the food traceability code associated with the task (such as the QR code on pork packaging), the full-chain node IDs bound to the traceability code in the supply chain node library are retrieved (such as PL-003 (live pig farming) → P-012 (slaughter and processing) → W-008 (frozen storage) → S-025 (supermarket terminal)). The target set of nodes for which data needs to be collected is then determined, which is the scope of nodes associated with the traceability task.
[0031] Dynamic correlation between two nodes The acquisition process is as follows: Label the two nodes of the associated edge as node i and node j respectively, and obtain the normalized geographical distance between node i and node j. (Get the straight-line distance between node i and node j) And obtain the maximum straight-line distance between nodes within the range of nodes associated with the tracing task. Through formula Calculate the normalized geographic distance ), obtain the normalized batch time difference between node i and node j. (Retrieve batch times for node i and node j) , (If node i is PL-003 and node j is P-012, then it is the time difference between the pig slaughter time of PL-003 and the slaughter time of P-012.) Get the maximum batch time difference allowed by the system by default. (System default), via formula Normalized batch time difference was calculated. ), obtain the degree centrality of node i Get node j Through formula ;in, Because the timeliness of batch time difference has a far greater impact on node linkages and risk transmission than geographical distance, a smaller batch time difference indicates more timely food circulation, more direct business linkages between nodes, and stronger risk transmission. The impact of geographical distance is weakened by factors such as cold chain logistics. The value can be 0.3. The value can be 0.7.
[0032] The degree centrality of a node is obtained as follows: Taking node i as an example, count the number of edges directly connected to node i in the network (including out-degree edges: edges originating from node i, such as farming → slaughtering; and in-degree edges: edges pointing to node i, such as feed → farming). This number is the original degree centrality of node i. The original degree centrality is normalized using the formula. The degree centrality of node i is calculated. N represents the total number of network nodes.
[0033] The high-probability initial risk screening index for a food safety traceability chain is obtained through the following steps: Select a food safety traceability chain, obtain the risk transmission consistency index between every two adjacent nodes in the chain, and sum and average all risk transmission consistency indices to calculate the average risk transmission consistency index. To obtain the timeliness deviation between every two adjacent nodes in the food safety traceability chain. The average timeliness deviation is calculated by summing and averaging all the timeliness deviations. Through formula The high-probability initial screening index for food safety traceability links was calculated. ;in, , All are weighting coefficients. Because the risk transmission consistency index is the core indicator for measuring whether the risk in the link has been truly transmitted—the abnormal coordination between upstream and downstream nodes (such as high veterinary drug residues upstream → excessive microorganisms downstream) is the essential causal relationship of the risk's existence and has a more critical impact. Timeliness deviation (Bave) is a catalytic factor for risk transmission. The value can be 0.6. The value can be 0.4.
[0034] The timeliness deviation between two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them node a and node b respectively, and obtain the actual processing time of node a and node b. Obtain the business attributes of node a and node b, and determine the standard transfer time based on the business attributes between node a and node b. Through formula The time difference between two adjacent nodes was calculated. .
[0035] The risk transmission consistency index between two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them node a and node b respectively. Obtain the detection index sets for nodes a and b. Filter out the detection data in the detection index set of node a that are associated with the food safety traceability task, and label them as the filtered index set. The detection data associated with the food safety traceability task in the detection index set of node b are selected and marked as the selected index set. (For example, in a cold chain pork spoilage risk traceability task, the detection index set for node P-012 (slaughtering and processing) includes: total microbial count (CFU / g), slaughter sterilization temperature, meat pH value, slaughtering time (min), and meat cutting weight (kg). Therefore, the screening index set includes total microbial count, slaughter sterilization temperature, and meat pH value, while slaughtering time and meat cutting weight are irrelevant data.) Obtain the business attributes of nodes a and b. Based on the business attributes between nodes a and b, determine multiple related index pairs from the two screening index sets. (For example, if the business attribute between nodes a and b is breeding → slaughtering, then from the two screening index sets, related index pairs such as veterinary drug residue ↔ total microbial count and pig body temperature ↔ slaughtering temperature can be screened. The definition of a related index pair is: association...) In an indicator pair, the upstream data being abnormal or normal can directly or indirectly lead to an increase or decrease in the risk of the downstream data, forming a transmission relationship of upstream cause → downstream risk result. The abnormal coherence index Pac for each associated indicator pair is obtained. The average of the abnormal coherence indices of all associated indicator pairs is calculated (or each associated indicator pair is assigned a weight, and the abnormal coherence index × weight is calculated, ensuring that the sum of the weights of all associated indicator pairs is 1; for example, veterinary drug residue → total microbial count is more critical than pig body temperature → slaughter temperature, therefore, the weight of veterinary drug residue → total microbial count is higher than the weight of pig body temperature → slaughter temperature; this will not be elaborated further here). The risk transmission consistency index between adjacent nodes is then calculated.
[0036] The abnormal consistency index Pac of the correlated indicator pair is obtained through the following steps: Select a correlated indicator pair, and obtain the upper and lower safety thresholds for the two test data in the pair (the upper and lower safety thresholds constitute the safety threshold range of the test data; each test data corresponds to one upper and one lower safety threshold, such as pH 6.0 to pH 7.0 for meat products, where pH 6.0 is the lower safety threshold and pH 7.0 is the upper safety threshold). Obtain the abnormal performance index API for each test data. Sum the abnormal performance indices of the two test data to calculate the abnormal performance superposition index PCA. Calculate the absolute difference between the abnormal performance indices of the two test data to obtain the abnormal performance difference index APD. Then, use the formula... The abnormal synergy index Pac of the associated indicator pair was calculated;
[0037] Abnormal performance index of detection data .
[0038] The food safety traceability and risk control module generates a core food safety traceability sequence based on the risk source link and non-source link. According to the core food safety traceability sequence, it sequentially traces high-risk candidate nodes and implements graded circuit breaker measures (such as hard circuit breaker (forced suspension) for the top 30% of high-risk candidate nodes in the core food safety traceability sequence, and conducts third-party re-inspection on all at-risk batches, upstream and downstream related batches, and non-source link batches sharing the node; soft circuit breaker (restriction of circulation) for the top 30%-70% of high-risk candidate nodes, and conducts sampling re-inspection on all at-risk batches of the node; and early warning circuit breaker (flexible monitoring) for the bottom 30% of high-risk candidate nodes, without forced suspension, only pushing risk warnings).
[0039] The core traceability nodes for food safety are determined based on risk source links and non-source links. The steps are as follows: Select risk source links, screen out high-risk candidate nodes in the risk source links, determine the core traceability index of each high-risk candidate node based on the non-source links, sort all high-risk candidate nodes in descending order of the core traceability index value, and generate a core traceability sequence for food safety based on the sorting.
[0040] The steps to screen out high-risk candidate nodes in the risk source link are as follows: obtain the source link contribution value of each node in the risk source link. When the source link contribution value is higher than the source link contribution threshold (the source link contribution threshold is set based on the statistical regularity of historical risk cases), the node is marked as a high-risk candidate node, and then all high-risk candidate nodes are counted.
[0041] The source link contribution value of a node is obtained through the following steps: Select a node, obtain the set of screening indicators for that node, and then obtain the anomaly performance index for each detection data in the set of screening indicators. Sum and average the anomaly performance indices of all detection data to calculate the anomaly performance contribution value Vfpc. Obtain the two risk transmission consistency indices of the two adjacent nodes (the risk transmission consistency index between this node and the upstream node, and the risk transmission consistency index between this node and the downstream node, respectively). Sum and average the two risk transmission consistency indices to calculate the adjacent anomaly contribution value ksd. Then, use the formula... The source link contribution value of this node is calculated. ;in, , All are weighting coefficients. Since node anomalies are the fundamental source of risk contribution, and adjacent risk transmission is a secondary factor amplifying risk, therefore... The value can be 0.7. The value can be 0.3.
[0042] The core traceability index for high-risk candidate nodes is determined by the following steps: Select a high-risk candidate node and obtain its degree centrality in the network topology. Each time a high-risk candidate node appears in a non-source link, the non-source link sharing frequency is incremented by one. Finally, the non-source link sharing frequencies are summed and marked as... The total number of non-source links is marked as M, and then calculated using the formula... The core traceability index of high-risk candidate nodes was calculated. ;in, , All are weighting coefficients. ,because This reflects the coverage breadth of high-risk candidate nodes across non-source links—the more non-source links a node appears on, the higher its probability of being a multi-link common-source risk point, and the more direct its impact on cross-link risks in the current tracing task. This is a core characteristic of core tracing nodes and should be assigned higher weight. Network connectivity strength is an auxiliary potential for risk diffusion; therefore, the value of z1 can be 0.6. The value can be 0.4.
[0043] The above formulas are all dimensionless calculations, and the preset parameters in the formulas should be set by those skilled in the art according to the actual situation.
[0044] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0045] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0046] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0047] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0048] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0049] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0050] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A food safety traceability and risk control system based on data mining, characterized in that, This includes a supply chain node library initialization module, a risk link marking module, and a food safety traceability and risk control module; The supply chain node library initialization module is used to initialize the supply chain node library and label the geographical coordinates, business attributes, and detection indicator sets of each node in the supply chain node library. The risk link marking module, upon receiving a food safety traceability task, locks down the range of nodes associated with the traceability task, designates all nodes within this range as network nodes, and identifies direct business relationships between nodes as network edges, forming a node-edge topology network. It then obtains the dynamic correlation degree of each associated edge in the topology network. When the dynamic correlation degree is higher than the dynamic correlation threshold and the nodes on both sides of the correlation edge are upstream and downstream business attributes, the correlation edge is marked as a core correlation edge. The batch time sequence of the core correlation edge is obtained. The core correlation edges are connected according to the batch time sequence to form multiple food safety traceability links. The high-probability initial screening index of each food safety traceability link is obtained. The food safety traceability link with the highest high-probability initial screening index is marked as the risk source link, and the rest of the food safety traceability links are marked as non-source links. The food safety traceability and risk control module generates a core food safety traceability sequence based on the risk source link and non-source link, and then performs traceability and graded circuit breaker measures on high-risk candidate nodes in sequence according to the core food safety traceability sequence.
2. The food safety traceability and risk control system based on data mining according to claim 1, characterized in that, After receiving a food safety traceability task, the scope of nodes associated with the traceability task is determined by the following steps: After receiving a food safety traceability task, the task tags are matched through the rule engine, and the preset task-node attribute mapping table is automatically associated. Through the food traceability code associated with the task, the full-chain node IDs bound to the traceability code in the supply chain node library are retrieved to determine the target set of nodes for which data needs to be collected, which is the scope of nodes associated with the traceability task.
3. The food safety traceability and risk control system based on data mining according to claim 1, characterized in that, Dynamic correlation between two nodes The acquisition process is as follows: Label the two nodes of the associated edge as node i and node j respectively, and obtain the normalized geographical distance between node i and node j. Obtain the normalized batch time difference between node i and node j. Obtain the degree centrality of node i. Get node j Through formula ;in, , All are weighting coefficients. .
4. A food safety traceability and risk control system based on data mining according to claim 1, characterized in that, The high-probability initial risk screening index for a food safety traceability chain is obtained through the following steps: Select a food safety traceability chain, obtain the risk transmission consistency index between every two adjacent nodes in the chain, and sum and average all risk transmission consistency indices to calculate the average risk transmission consistency index. To obtain the timeliness deviation between every two adjacent nodes in the food safety traceability chain. The average timeliness deviation is calculated by summing and averaging all the timeliness deviations. Through formula The high-probability initial screening index of the food safety traceability chain was calculated. ;in, , All are weighting coefficients. .
5. A food safety traceability and risk control system based on data mining according to claim 4, characterized in that, The timeliness deviation between two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them node a and node b respectively, and obtain the actual processing time of node a and node b. Obtain the business attributes of node a and node b, and determine the standard transfer time based on the business attributes between node a and node b. Through formula The time difference between two adjacent nodes was calculated. .
6. A food safety traceability and risk control system based on data mining according to claim 4, characterized in that, The risk transmission consistency index between two adjacent nodes is obtained through the following steps: Select two adjacent nodes in the food safety traceability chain, labeling them node a and node b respectively. Obtain the detection index sets for nodes a and b. Filter out the detection data in the detection index set of node a that are associated with the food safety traceability task, and label them as the filtered index set. The detection data associated with the food safety traceability task in the detection index set of node b are selected and marked as the selected index set. The business attributes of node a and node b are obtained. Based on the business attributes between node a and node b, multiple related indicator pairs are determined from two sets of screening indicators. The abnormal coordination consistency index Pac of each related indicator pair is obtained. The average of the abnormal coordination consistency indices of all related indicator pairs is summed to calculate the risk transmission consistency index of two adjacent nodes.
7. A food safety traceability and risk control system based on data mining according to claim 6, characterized in that, The abnormal consistency index Pac of a correlated indicator pair is obtained through the following steps: Select a correlated indicator pair, obtain the upper and lower safety thresholds for the two detection data in the pair, obtain the abnormal performance index API for each detection data, sum the abnormal performance indices of the two detection data to calculate the abnormal performance superposition index PCA, and calculate the absolute difference between the abnormal performance indices of the two detection data to obtain the abnormal performance difference index APD. The abnormal synergy index Pac of the associated index pair is calculated.
8. A food safety traceability and risk control system based on data mining according to claim 1, characterized in that, The core traceability nodes for food safety are determined based on risk source links and non-source links. The steps are as follows: Select risk source links, screen out high-risk candidate nodes in the risk source links, determine the core traceability index of each high-risk candidate node based on the non-source links, sort all high-risk candidate nodes in descending order of the core traceability index value, and generate a core traceability sequence for food safety based on the sorting.
9. A food safety traceability and risk control system based on data mining according to claim 8, characterized in that, The steps to screen out high-risk candidate nodes in the risk source link are as follows: obtain the source link contribution value of each node in the risk source link. When the source link contribution value is higher than the source link contribution threshold, the node is marked as a high-risk candidate node. Then, all high-risk candidate nodes are counted. The source link contribution value of a node is obtained through the following steps: Select a node, obtain the set of screening indicators for that node, and then obtain the anomaly performance index for each detection data in the set of screening indicators. Sum and average the anomaly performance indices of all detection data to calculate the anomaly performance contribution value Vfpc. Obtain the two risk transmission consistency indices of the two adjacent nodes, sum and average the two risk transmission consistency indices to calculate the adjacent anomaly contribution value ksd. Then, use the formula... The source link contribution value of this node is calculated. ;in, , All are weighting coefficients. .
10. A food safety traceability and risk control system based on data mining according to claim 8, characterized in that, The core traceability index for high-risk candidate nodes is determined by the following steps: Select a high-risk candidate node and obtain its degree centrality in the network topology. Each time a high-risk candidate node appears in a non-source link, the non-source link sharing frequency is incremented by one. Finally, the non-source link sharing frequencies are summed and marked as... The total number of non-source links is marked as M, and then calculated using the formula... The core traceability index of high-risk candidate nodes was calculated. ;in, , All are weighting coefficients. .