Supply chain supplier accurate matching method and system based on big data analysis

Abstract

The invention provides a supply chain supplier accurate matching method and system based on big data analysis. Firstly, a supply chain field big data set is constructed; constructing a supply chain knowledge graph based on the big data set, and identifying and quantifying a deviation risk level of a current purchase demand relative to a historical purchase reference mode in the knowledge graph; and calculating a supply network stability index and a material replacement difficulty index of the material in the current purchase list. And finally, performing judgment according to the deviation risk level, a supply network stability index and a material replacement difficulty index, and dynamically selecting a supplier matching strategy: if the overall risk is controllable, the supply network is stable and the material replacement difficulty is low, starting multi-dimensional similarity matching based on the whole knowledge graph to perform global supplier matching; and otherwise, triggering a credible supplier priority recommendation mechanism based on the historical cooperation subgraph and the performance big data. According to the scheme, accurate matching of supply chain suppliers can be realized.

Description

Technical Field

[0001] This application relates to the field of supply chain management technology, and more specifically, to a method and system for precise matching of supply chain suppliers based on big data analysis. Background Technology

[0002] In supply chain management, supplier matching is a core element in ensuring procurement efficiency, cost control, and supply chain stability. With the deepening of globalization and increased market volatility, enterprises face challenges such as diversified procurement needs, fragmented supplier resources, and uncertainties in risk. Traditional supplier matching primarily relies on manual screening or simple database queries. This method selects partners by comparing basic supplier qualifications and historical quotes, but it ignores the dynamic interconnectedness of the supply chain and the potential value of big data, resulting in low matching accuracy and an inability to effectively respond to sudden changes in demand or supply chain disruptions. Therefore, achieving accurate supplier matching has become a pressing issue for the industry. Summary of the Invention

[0003] This application provides a method and system for precise matching of supply chain suppliers based on big data analysis, which can achieve precise matching of supply chain suppliers.

[0004] Firstly, this application provides a method for precise matching of supply chain suppliers based on big data analysis, including: Integrate procurement needs, supplier profiles, material master data, and historical transaction records to build a big data collection in the supply chain field; A supply chain knowledge graph is constructed based on the big data collection in the supply chain field, where nodes include supplier entities, material entities, and purchaser entities, and edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data. Identify and quantify the risk level of deviation of current procurement demand from historical procurement benchmark patterns in the supply chain knowledge graph; By mining the topology of the supply chain knowledge graph and analyzing the association rules of historical transaction data, the supply network stability index and the material substitution difficulty index of the materials in the current procurement list are calculated. Based on the deviation risk level, the supply network stability index, and the material substitution difficulty index, a supplier matching strategy is dynamically selected: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then a global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

[0005] In some embodiments, the historical procurement benchmark model is a benchmark model formed by aggregating historical procurement data according to material categories based on the big data set in the supply chain field. The benchmark model includes at least procurement quantity benchmark, delivery cycle benchmark, and cost benchmark.

[0006] In some embodiments, identifying and quantifying the risk level of deviation of current procurement demand from historical procurement benchmark patterns in the supply chain knowledge graph specifically includes: Based on the material specifications, purchase quantity, delivery cycle and budget cost information recorded in the supply chain knowledge graph, the deviation values ​​between the current purchase demand and the corresponding historical purchase benchmark model are calculated respectively. Assign a preset weight coefficient to each deviation dimension; The weighted composite deviation value is calculated based on the deviation values ​​and weighting coefficients of each dimension; Based on a preset deviation threshold range, the weighted comprehensive deviation value is mapped to a discrete deviation risk level.

[0007] In some embodiments, the calculation of the supply network stability index specifically includes: Based on the topology of the supply chain knowledge graph, the topological features of the supplier nodes associated with the current material are mined and extracted, including the degree centrality and network dependency of the supplier nodes. By combining association rule analysis with historical transaction data, we can calculate the supplier's capacity volatility, historical on-time delivery rate, and financial status score. By integrating the topological characteristics with the capacity volatility, historical on-time delivery rate, and financial status score, a quantitative supply network stability index is calculated using a preset stability assessment model.

[0008] In some embodiments, the calculation of the material substitution difficulty index specifically includes: Based on the material master data and relationships in the supply chain knowledge graph, the matching degree of technical parameters between the candidate alternative materials and the original materials in terms of key performance parameters and specifications is calculated. Based on the association rule analysis of supplier profiles and historical transaction records in the supply chain knowledge graph, the difficulty and cost differences in obtaining the required certification qualifications for candidate alternative materials are assessed. Taking into account the matching degree of the technical parameters, the difficulty of obtaining certification qualifications, and the cost differences, a quantitative material substitution difficulty index is calculated through a preset substitution difficulty assessment model.

[0009] In some embodiments, the criteria for judging whether the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low in the dynamic supplier selection matching strategy are as follows: The deviation risk level is lower than a preset first risk threshold; The stability index of the supply network is higher than the preset stability threshold; The material substitution difficulty index is lower than the preset substitution difficulty threshold; When all three conditions above are met, global supplier matching based on the entire supply chain knowledge graph is initiated; otherwise, a trusted supplier priority recommendation mechanism is triggered.

[0010] In some embodiments, global supplier matching is performed based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph, specifically including: Traverse the supply chain knowledge graph and initialize all supplier-type nodes as a global candidate node set; For each supplier node in the global candidate node set, featureization and matching degree calculation based on the graph structure are performed, wherein: Read the attribute data of the supplier node itself and calculate the supplier qualification matching degree according to the preset qualification assessment rules; Traverse the historical transaction relationship edges between the supplier node and the current purchaser node, extract the historical transaction evaluation data carried by the relationship edges, and calculate the historical cooperation performance matching degree according to the preset performance evaluation rules; Traverse the supply relationship edges between the supplier node and the material nodes in the current purchase list, extract the supply capacity data carried by the relationship edges, and calculate the material supply capacity matching degree according to the preset supply capacity evaluation rules; Based on the preset evaluation rules, the matching degree of supplier qualifications, historical cooperation performance, and material supply capacity of each supplier node are evaluated to generate a global matching score for that supplier node. Based on the global matching score of all supplier nodes, the global candidate node set is sorted and a supplier recommendation list is output.

[0011] Secondly, this application provides a supply chain supplier precision matching system based on big data analysis, which includes: The module is used to integrate procurement needs, supplier profiles, material master data, and historical transaction records to build a big data collection in the supply chain field. The processing module is used to construct a supply chain knowledge graph based on the big data set in the supply chain field, wherein the nodes include supplier entities, material entities, and purchaser entities, and the edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data; The processing module is also used to identify and quantify the deviation risk level of the current procurement demand relative to the historical procurement benchmark pattern in the supply chain knowledge graph. The processing module is also used to calculate the supply network stability index and material substitution difficulty index of the materials in the current procurement list by mining the topology of the supply chain knowledge graph and analyzing the association rules of historical transaction data. The execution module is used to make judgments based on the deviation risk level, the supply network stability index, and the material substitution difficulty index, and dynamically select a supplier matching strategy: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

[0012] Thirdly, this application provides a computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described supply chain supplier precision matching method based on big data analysis.

[0013] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for precise matching of supply chain suppliers based on big data analysis.

[0014] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In the supply chain supplier matching process, this application achieves multi-source data fusion of procurement needs, supplier profiles, material master data, and historical transaction records by constructing a supply chain knowledge graph. By identifying deviation risk levels, supply network stability indicators, and material substitution difficulty indicators, it dynamically selects matching strategies, avoiding the limitations of traditional methods. The global matching strategy improves matching accuracy by utilizing the multi-dimensional similarity of the knowledge graph, while the trusted supplier priority mechanism ensures the reliability of the supply chain in high-risk scenarios, ultimately achieving accurate supplier matching and providing efficient support for enterprise procurement decisions. Attached Figure Description

[0015] Figure 1 This is an exemplary flowchart illustrating a supply chain supplier precision matching method based on big data analysis, according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the identification and quantification of deviation risk levels according to some embodiments of this application; Figure 3 This is a schematic diagram of the structure of a supply chain supplier precision matching system based on big data analysis, according to some embodiments of this application; Figure 4This is a schematic diagram of the structure of a computer device for implementing a supply chain supplier precision matching method based on big data analysis, according to some embodiments of this application. Detailed Implementation

[0016] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific implementation methods. (Reference) Figure 1 The figure is an exemplary flowchart of a supply chain supplier precision matching method based on big data analysis, according to some embodiments of this application. The method mainly includes the following steps: In step 101, procurement requirements, supplier profiles, material master data, and historical transaction records are integrated to construct a big data collection in the supply chain field.

[0017] The procurement requirements include information such as material specifications, quantities, delivery cycles, and budgeted costs in the current procurement list; supplier profiles include static data such as supplier qualifications, production capacity, financial status, and geographical location; material master data includes material specifications, performance parameters, and substitution relationships; and historical transaction records include dynamic data such as on-time delivery rates, quality evaluations, and cost fluctuations from past transactions. In some embodiments, the construction of the supply chain big data set can be specifically carried out in the following ways: First, integrate multi-source data, namely, collect four core data categories: procurement requirements (including material specifications, quantities, delivery cycles, budgets, etc., data can come from ERP systems), supplier profiles (including qualifications, capacity, financial status, etc., data can come from CRM systems), material master data (including specifications, performance parameters, substitution relationships, etc.), and historical transaction records (including on-time delivery rate, quality evaluation, cost fluctuations, etc., from transaction logs). Then, perform data preprocessing on the integrated multi-source data, for example, using ETL tools (such as Apache NiFi) to clean, deduplicate, and standardize the integrated multi-source data to eliminate data inconsistencies. Finally, unify the storage, for example, through a distributed storage system, to aggregate the preprocessed structured / semi-structured data to form a unified big data set in the supply chain field. This is only an example and is not intended to limit the scope of this application.

[0018] In step 102, a supply chain knowledge graph is constructed based on the big data set in the supply chain field. The nodes include supplier entities, material entities, and purchaser entities, while the edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data.

[0019] In practical implementation, building a supply chain knowledge graph can be done in the following way: First, data preprocessing is performed, which involves filtering and organizing core information about suppliers, materials, and buyers from the existing big data collection in the supply chain field. This includes supplier qualifications, material specifications, buyer requirements, and various related data such as supply records and transaction evaluations, ensuring that the data is structured and usable. Then, entity and relationship extraction is performed. For node construction, entity recognition algorithms (such as NER) can be used to extract specific instances of supplier, material, and buyer entities from the core information data, serving as the three core nodes of the knowledge graph, such as supplier A, material X, and buyer B. Edge construction can be achieved through association rule mining algorithms (such as FP). -Growth) identifies relationships between nodes from various types of related data, such as supply relationships mapped between suppliers and materials, cooperation history mapped between suppliers and buyers, and material matching relationships mapped between materials. At the same time, transaction evaluation data is attached to the corresponding edges. Finally, graph storage is performed, that is, the extracted nodes, relationships and edge-carried data are loaded into a graph database (such as Neo4j) to form a complete supply chain knowledge graph, realizing the visualized and associated storage of nodes and relationships. The nodes include supplier entities, material entities, and buyer entities, and the edges represent supply relationships, cooperation history, material matching relationships and transaction evaluation data, which will not be elaborated here.

[0020] In step 103, the deviation risk level of the current procurement demand relative to the historical procurement benchmark pattern is identified and quantified in the supply chain knowledge graph.

[0021] The historical procurement benchmark model is a benchmark model formed by aggregating historical procurement data according to material categories based on the big data set in the supply chain field. This benchmark model includes at least procurement quantity benchmarks, delivery cycle benchmarks, and cost benchmarks. In some embodiments, reference is made to... Figure 2 This diagram is an exemplary flowchart for identifying and quantifying the level of deviation risk: In step 1031, based on the material specifications, purchase quantity, delivery cycle, and budget cost information recorded in the supply chain knowledge graph, the deviation values ​​between the current purchase demand and the corresponding historical purchase benchmark patterns are calculated. Specifically, for example, the four core data categories of the current purchase demand (i.e., key parameters of material specifications, purchase quantity, delivery cycle, and budget cost) and the benchmark values ​​of these four categories of data from the historical purchase benchmark patterns of the corresponding materials are first extracted from the supply chain knowledge graph. The deviation is calculated separately for each dimension. The purchase quantity, delivery cycle, and budget cost can be calculated using relative deviation formulas, such as |current value - benchmark value| / benchmark value) or Euclidean distance. The deviation value for material specifications can be obtained by comparing the differences in key performance / size parameters. Finally, the independent deviation values ​​corresponding to the material specifications, purchase quantity, delivery cycle, and budget cost are obtained, without comprehensive calculation. In step 1032, a preset weight coefficient is assigned to each deviation dimension. For example, the purchase quantity weight of standardized materials (such as standard parts) is 0.3, the delivery cycle weight is 0.2, the cost weight is 0.2, and the material specification weight is 0.3. The purchase quantity weight of customized materials (such as special equipment parts) is 0.2, the delivery cycle weight is 0.1, the cost weight is 0.1, and the material specification weight is 0.6. The total weight is 1. The weights can be dynamically adjusted according to the material type. This is only an example and is not intended to limit the specific scope of this application. In step 1033, the weighted comprehensive deviation value is calculated based on the deviation values ​​and weighting coefficients of each dimension. For example, the independent deviation values ​​of four dimensions—material specifications, purchase quantity, delivery cycle, and budgeted cost—have been calculated, and preset weighting coefficients have been assigned to each dimension. For example, for standardized materials: purchase quantity 0.3, delivery cycle 0.2, cost 0.2, material specifications 0.3; for customized materials: purchase quantity 0.2, delivery cycle 0.1, cost 0.1, material specifications 0.6. Then, the weighted deviation of each dimension can be calculated, that is, the deviation value of each dimension is multiplied by the corresponding weighting coefficient. Finally, the result is obtained by summing all the weighted deviations of each dimension, and the final sum is the weighted comprehensive deviation value. In step 1034, based on a preset deviation threshold range (e.g., 0-0.2 for low risk, 0.2-0.5 for medium risk, and >0.5 for high risk), the weighted comprehensive deviation value is mapped to a discrete deviation risk level. In specific implementation, fixed deviation threshold ranges and corresponding discrete risk levels can be defined in advance, for example: 0-0.2 for low risk, 0.2-0.5 for medium risk, and >0.5 for high risk. The specific range can be adjusted according to industry characteristics. Then, the previously calculated weighted comprehensive deviation value is retrieved; the relationship between the weighted comprehensive deviation value and each threshold range is compared; and it is mapped to the discrete risk level of the corresponding range. For example, a comprehensive deviation value of 0.3 represents medium risk. This is only an example and is not intended to limit the scope of this application.

[0022] It should be noted that the deviation risk level in this application can be dynamically adjusted based on industry characteristics. For example, the manufacturing industry pays more attention to delivery cycle, which will not be elaborated here.

[0023] In step 104, the supply network stability index and material substitution difficulty index of the materials in the current procurement list are calculated by mining the topology of the supply chain knowledge graph and analyzing the association rules of historical transaction data.

[0024] In some embodiments, the supply network stability index may be calculated in the following manner: Based on the topology of the supply chain knowledge graph, the topological features of the supplier nodes associated with the current material are mined and extracted. The topological features include the degree centrality of the supplier nodes (i.e., the number of direct connection edges between the supplier node and all other nodes in the supply chain knowledge graph (including material nodes, purchaser nodes, and other supplier nodes)) and network dependency (i.e., path dependency). The topology of the supply chain knowledge graph refers to the connection pattern between nodes, such as the relationship between suppliers and materials, and other suppliers. The degree centrality is measured by the total number of connected related nodes. The more connected nodes, the more core the supplier is in the supply network. Network dependency is based on path dependency. Specifically, it refers to whether the path to the supplier is unique or singular during the current material acquisition process. The more singular the path, the higher the dependency; the more diverse the path (which can be obtained through connections with other nodes), the lower the dependency. By combining association rule analysis with historical transaction data, the following metrics are calculated for suppliers: capacity volatility (calculated based on monthly capacity data for the past 12 months, using the formula: monthly capacity standard deviation / monthly capacity mean; if historical data is less than 12 months, the actual usable data is used and data integrity is noted), historical delivery timeliness, and financial status score. Specifically, through association rule analysis, effective correlation information related to supplier capacity, delivery, and finance can be extracted from historical transaction data, providing data support for indicator calculation. Capacity volatility can be obtained by dividing the standard deviation of the supplier's historical capacity data by the mean. The standard deviation reflects the amplitude of capacity fluctuation, while the mean is the average capacity; a higher ratio indicates more unstable capacity. Historical delivery timeliness can be calculated by statistically analyzing the proportion of on-time deliveries in historical transactions to the total number of deliveries, directly reflecting delivery reliability. The financial status score is based on a preset credit scoring model, inputting financial data from the supplier's historical transactions, such as payment collection speed and cash flow, and the model calculates a quantitative score reflecting the supplier's financial health. Further details are omitted here. By integrating the topological features with the capacity volatility, historical on-time delivery rate and financial status score, a quantitative supply network stability index is calculated through a preset stability assessment model (this model is a weighted linear regression model, trained based on nearly 3 years of supply chain historical data (including 2000+ stable cases and 500+ interruption cases), with a verified accuracy of ≥85% and recall of ≥80%, and the model parameters can be iteratively optimized periodically).

[0025] The supply network stability index is used to measure the reliability of the material supply network. The higher the value, the more stable the supply network. In some embodiments, the topological features are integrated with the capacity volatility, historical delivery timeliness and financial status score. The quantitative supply network stability index is calculated by a preset stability assessment model. Specifically, the following method can be used: First, the topological features (i.e., degree centrality, network dependence) and the operational indicators (i.e., capacity volatility, historical delivery timeliness and financial status score) are unified in terms of dimensions and transformed into standardized values ​​in the 0-1 range (eliminating the influence of differences in the units of different indicators), that is, data standardization. Then, weight coefficients are assigned, that is, preset weights are assigned to each standardized feature / indicator. For example, the degree centrality weight is 0.2, the network dependency weight is 0.2, the capacity volatility weight is 0.2, the on-time delivery rate weight is 0.3, and the financial score weight is 0.1, with a total weight of 1. These weights can be adjusted according to actual industry needs. Then, the standardized values ​​of each feature / indicator are multiplied by their corresponding weights, and then summed through a preset model to obtain the final integrated quantitative result, which serves as the core calculation basis for the supply network stability index. The preset model is used to adapt to the integration of multiple data in business scenarios. For example, it can be a weighted linear regression model. This is only an example and is not intended to limit the specific scope of this application.

[0026] In addition, the material substitution difficulty index is used to measure the ease with which a candidate alternative material can replace the original material. In some embodiments, the material substitution difficulty index can be calculated in the following ways: Based on the material master data and relationships in the supply chain knowledge graph, the matching degree of technical parameters between candidate substitute materials and original materials in terms of key performance parameters and specifications is calculated. Specifically, for example, key performance parameters (e.g., material, pressure resistance) and specifications (e.g., length, width, height, interface type) of the original materials are extracted from the material master data in the supply chain knowledge graph to form a set of original material parameters. Then, through the substitution / matching relationships between materials in the knowledge graph, candidate substitute materials are selected, and their corresponding key performance parameters and specifications are extracted simultaneously to form a set of candidate material parameters. Finally, the parameter sets of the original materials and candidate substitute materials are compared to calculate their overlap / similarity, ultimately obtaining a quantified technical parameter matching degree (value 0-1, higher values ​​indicate better matching). The calculation of overlap / similarity between the parameter sets of the original materials and candidate substitute materials can use algorithms such as cosine similarity; this is merely an example and not intended to limit the scope of this application. Based on the association rule analysis of supplier profiles and historical transaction records in the supply chain knowledge graph, the difficulty of obtaining the required certification qualifications for candidate alternative materials (including average certification time, time fluctuation value, and certification pass rate) and cost differences are assessed. Specifically, this involves extracting the certification qualification requirements (e.g., industry certification, quality standards) corresponding to candidate alternative materials from the supplier profiles in the supply chain knowledge graph, and retrieving certification-related data for similar materials (including historical certification time, certification fees, procurement costs, and certification pass rate) from historical transaction records. Then, association rule analysis is performed to find the relationships between material type, certification qualifications, certification time, certification pass rate, and cost, and to filter historical data samples matching the current candidate alternative materials. This association rule analysis can be performed using association rule mining algorithms (e.g., FP-Growth) or other association rule analysis algorithms; no specific limitations are made here. Finally, the difficulty of obtaining certification is assessed by calculating the average certification time and time fluctuation value based on historical certification time samples, combined with the certification pass rate, using the formula: "Certification acquisition difficulty = (average certification time / industry average time) × 0.4 + (time fluctuation value / industry average fluctuation value) × 0.3". The result is calculated as "+(1-certification pass rate)×0.3", and standardized to a range of 0-1. The longer the time, the greater the fluctuation, and the lower the pass rate, the more difficult it is to obtain certification. Finally, the cost difference is calculated by comparing the historical procurement costs of the original materials and the candidate alternative materials, as well as the certification-related expenses (testing fees, audit fees), to determine the cost difference or cost ratio between the two and clarify the size of the cost difference. Based on the combined factors of technical parameter matching degree, certification difficulty, and cost difference, a quantified material substitution difficulty index (range 0-1, with low values ​​indicating easy substitution) is calculated using a pre-defined substitution difficulty assessment model (e.g., fuzzy logic model). Specifically, for example, the technical parameter matching degree, certification difficulty, and cost difference are standardized to a 0-1 range; simultaneously, the direction of the index is adjusted, with higher technical matching degree indicating easier substitution, and higher certification difficulty and cost difference indicating more difficult substitution, ensuring consistent impact of the three factors on substitution difficulty. Then, rules are pre-defined in the fuzzy logic model to cover different index combinations; for example, low technical matching degree + high certification difficulty + large cost difference indicates high substitution difficulty, while high technical matching degree + low certification difficulty + small cost difference indicates low substitution difficulty. The pre-processed three indicators are input into the substitution difficulty assessment model, and fuzzy reasoning is performed according to the pre-defined rules to obtain high / medium / low fuzzy results. Finally, defuzzification processing, such as the centroid method, is used to convert the fuzzy results into quantified values ​​in the 0-1 range, which are the material substitution difficulty indexes. The lower the value of this index, the easier the substitution.

[0027] In step 105, a supplier matching strategy is dynamically selected based on the deviation risk level, the supply network stability index, and the material substitution difficulty index: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then a global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

[0028] The supplier matching strategy includes global supplier matching and a trusted supplier priority recommendation mechanism. Global suppliers refer to a set of candidate suppliers selected based on multi-dimensional data across the entire supply chain. This differs from local suppliers selected based on single criteria such as price or region. The evaluation covers all aspects, including qualification compliance, historical cooperation performance, supply network stability, and material and technology adaptability. This comprehensive approach can meet the procurement needs and support the long-term stable operation of the supply chain.

[0029] The trusted supplier priority recommendation mechanism refers to a strategy that, in procurement scenarios with high overall risk levels, unstable supply networks, or high difficulty in material substitution, selects trusted suppliers with good credit, reliable delivery, and controllable risks based on supply chain knowledge graphs and prioritizes their inclusion in the procurement candidate list in order to reduce supply chain risks and ensure the stable implementation of procurement needs.

[0030] It should be noted that the criteria for judging the overall risk as controllable, the supply network as stable, and the material substitution difficulty as low are as follows: the deviation risk level is lower than a preset first risk threshold (e.g., below medium risk); the supply network stability index is higher than a preset stability threshold (e.g., above 0.7); and the material substitution difficulty index is lower than a preset substitution difficulty threshold (e.g., below 0.3). When all three conditions are met, global supplier matching based on the entire supply chain knowledge graph is initiated. Conversely, if any one of the above three conditions is not met, it can be determined that the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high.

[0031] In some embodiments, global supplier matching is performed based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph. Specifically, the following methods can be used: Traverse the supply chain knowledge graph and initialize all supplier type nodes as a global candidate node set. For each supplier node in the global candidate node set, perform feature mapping and matching degree calculation based on the graph structure, including: reading the attribute data of the supplier node itself and calculating the supplier qualification matching degree according to preset qualification assessment rules (such as cosine similarity matching qualification requirements); traversing the historical transaction relationship edges between the supplier node and the current purchaser node, extracting the historical transaction evaluation data carried by the relationship edges, and calculating the historical cooperation performance matching degree according to preset performance evaluation rules (such as weighted average score); traversing the supply relationship edges between the supplier node and the material nodes in the current purchase list, extracting the supply capacity data carried by the relationship edges, and calculating the supply capacity matching degree according to preset supply capacity assessment rules (such as capacity coverage). The matching degree of material supply capacity is calculated using the coverage rate. Based on preset evaluation rules (e.g., weighted summation), weights are assigned to suit different procurement scenarios: In emergency procurement scenarios, supplier qualification weight is 0.2, historical cooperation performance weight is 0.2, and material supply capacity weight is 0.6; in long-term strategic cooperation scenarios, supplier qualification weight is 0.4, historical cooperation performance weight is 0.4, and material supply capacity weight is 0.2; in regular standardized procurement scenarios, supplier qualification weight is 0.3, historical cooperation performance weight is 0.3, and material supply capacity weight is 0.4. The matching degree of supplier qualification, historical cooperation performance, and material supply capacity for each supplier node is evaluated to generate a global matching score for that supplier node. Based on the global matching scores of all supplier nodes, the global candidate node set is sorted, and a supplier recommendation list is output.

[0032] In some embodiments, triggering the trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data can be achieved in the following ways: The system is automatically triggered when any one of the following conditions is met: a high overall risk level, an unstable supply network, or high difficulty in material substitution. It extracts historical cooperation subgraphs from the supply chain knowledge graph. These subgraphs contain topological information such as historical cooperation links, frequency of cooperation, and product category relationships between suppliers that belong to the same category (same type) according to GB / T 7635.1 material classification codes and use the same core raw materials and industry standards (same source) as the currently procured materials. Simultaneously, it retrieves supplier performance big data, covering core performance indicators such as historical on-time delivery rate, quality pass rate, contract performance stability, and emergency response capability. Based on the closeness of the historical cooperation subgraphs and the quantitative scoring of performance big data, the system ranks candidate suppliers by credibility, prioritizing suppliers with good credit, excellent performance, and stable cooperation in the procurement candidate list and outputting a recommendation priority.

[0033] It should be noted that the historical cooperation subgraph in this application is a partial subgraph of the supply chain knowledge graph. It consists of entities and relationships between entities related to historical procurement of materials of the same type or origin as the target procurement materials. It is used to accurately extract the historical cooperation characteristics of suppliers and can provide visualized topological and attribute information such as historical cooperation links, cooperation frequency, and cooperation stability for the screening of trustworthy suppliers.

[0034] Furthermore, in another aspect of this application, in some embodiments, this application provides a supply chain supplier precision matching system based on big data analysis, referencing... Figure 3 The figure is a schematic diagram of the structure of a supply chain supplier precision matching system based on big data analysis, according to some embodiments of this application. The system 300 includes: a construction module 301, a processing module 302, and an execution module 303, which are described below: Module 301 in this application is mainly used to integrate procurement needs, supplier profiles, material master data and historical transaction records to build a big data set in the supply chain field. Processing module 302 in this application is mainly used to construct a supply chain knowledge graph based on the big data set in the supply chain field. The nodes include supplier entities, material entities, and purchaser entities, and the edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data. Processing module 302 is also used to identify and quantify the deviation risk level of the current procurement demand relative to the historical procurement benchmark pattern in the supply chain knowledge graph. Processing module 302 is also used to calculate the supply network stability index and material substitution difficulty index of the materials in the current procurement list by mining the topology structure of the supply chain knowledge graph and analyzing the association rules of historical transaction data. The execution module 303 in this application is mainly used to make judgments based on the deviation risk level, the supply network stability index, and the material substitution difficulty index, and dynamically select a supplier matching strategy: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then a global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

[0035] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described supply chain supplier precision matching method based on big data analysis.

[0036] In some embodiments, reference Figure 4The figure is a schematic diagram of the structure of a computer device implementing a supply chain supplier precision matching method based on big data analysis, according to some embodiments of this application. The methods in the above embodiments can... Figure 4 The computer device shown is used to implement this, and the computer device 400 includes at least one processor 401, a communication bus 402, a memory 403, and at least one communication interface 404.

[0037] The processor 401 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more for controlling the execution of the supply chain supplier precision matching method based on big data analysis in this application.

[0038] The communication bus 402 may include a path for transmitting information between the aforementioned components.

[0039] The memory 403 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or it may be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CDROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory 403 may exist independently and be connected to the processor 401 via a communication bus 402. The memory 403 may also be integrated with the processor 401.

[0040] The memory 403 stores program code for executing the scheme of this application, and its execution is controlled by the processor 401. The processor 401 executes the program code stored in the memory 403. The program code may include one or more software modules. In the above embodiments, the quantification of the deviation risk level can be achieved by the processor 401 and one or more software modules in the program code in the memory 403.

[0041] Communication interface 404 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0042] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single CPU) processor or a multi-core (multi CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0043] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0044] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for precise matching of supply chain suppliers based on big data analysis.

[0045] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0046] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for precise matching of supply chain suppliers based on big data analysis, characterized in that, include: Integrate procurement needs, supplier profiles, material master data, and historical transaction records to build a big data collection in the supply chain field; A supply chain knowledge graph is constructed based on the big data collection in the supply chain field, where nodes include supplier entities, material entities, and purchaser entities, and edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data. Identify and quantify the risk level of deviation of current procurement demand from historical procurement benchmark patterns in the supply chain knowledge graph; By mining the topology of the supply chain knowledge graph and analyzing the association rules of historical transaction data, the supply network stability index and the material substitution difficulty index of the materials in the current procurement list are calculated. Based on the deviation risk level, the supply network stability index, and the material substitution difficulty index, a supplier matching strategy is dynamically selected: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then a global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

2. The method according to claim 1, characterized in that, The historical procurement benchmark model is a benchmark model formed by aggregating historical procurement data according to material categories based on the big data set in the supply chain field. The benchmark model includes at least procurement quantity benchmark, delivery cycle benchmark and cost benchmark.

3. The method according to claim 1, characterized in that, Identifying and quantifying the risk level of deviation between current procurement needs and historical procurement benchmark patterns within the aforementioned supply chain knowledge graph specifically includes: Based on the material specifications, purchase quantity, delivery cycle and budget cost information recorded in the supply chain knowledge graph, the deviation values ​​between the current purchase demand and the corresponding historical purchase benchmark model are calculated respectively. Assign a preset weight coefficient to each deviation dimension; The weighted composite deviation value is calculated based on the deviation values ​​and weighting coefficients of each dimension; Based on a preset deviation threshold range, the weighted comprehensive deviation value is mapped to a discrete deviation risk level.

4. The method according to claim 1, characterized in that, The calculation of the supply network stability index specifically includes: Based on the topology of the supply chain knowledge graph, the topological features of the supplier nodes associated with the current material are mined and extracted, including the degree centrality and network dependency of the supplier nodes. By combining association rule analysis with historical transaction data, we can calculate the supplier's capacity volatility, historical on-time delivery rate, and financial status score. By integrating the topological characteristics with the capacity volatility, historical on-time delivery rate, and financial status score, a quantitative supply network stability index is calculated using a preset stability assessment model.

5. The method according to claim 1, characterized in that, The calculation of the material substitution difficulty index specifically includes: Based on the material master data and relationships in the supply chain knowledge graph, the matching degree of technical parameters between the candidate alternative materials and the original materials in terms of key performance parameters and specifications is calculated. Based on the association rule analysis of supplier profiles and historical transaction records in the supply chain knowledge graph, the difficulty and cost differences in obtaining the required certification qualifications for candidate alternative materials are assessed. Taking into account the matching degree of the technical parameters, the difficulty of obtaining certification qualifications, and the cost differences, a quantitative material substitution difficulty index is calculated through a preset substitution difficulty assessment model.

6. The method according to claim 1, characterized in that, In the aforementioned dynamic supplier selection and matching strategy, the criteria for judging overall risk controllability, supply network stability, and low difficulty in material substitution are as follows: The deviation risk level is lower than a preset first risk threshold; The stability index of the supply network is higher than the preset stability threshold; The material substitution difficulty index is lower than the preset substitution difficulty threshold; When all three conditions above are met, global supplier matching based on the entire supply chain knowledge graph is initiated; otherwise, a trusted supplier priority recommendation mechanism is triggered.

7. The method according to claim 1, characterized in that, Global supplier matching based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph specifically includes: Traverse the supply chain knowledge graph and initialize all supplier-type nodes as a global candidate node set; For each supplier node in the global candidate node set, featureization and matching degree calculation based on the graph structure are performed, wherein: Read the attribute data of the supplier node itself and calculate the supplier qualification matching degree according to the preset qualification assessment rules; Traverse the historical transaction relationship edges between the supplier node and the current purchaser node, extract the historical transaction evaluation data carried by the relationship edges, and calculate the historical cooperation performance matching degree according to the preset performance evaluation rules; Traverse the supply relationship edges between the supplier node and the material nodes in the current purchase list, extract the supply capacity data carried by the relationship edges, and calculate the material supply capacity matching degree according to the preset supply capacity evaluation rules; Based on the preset evaluation rules, the matching degree of supplier qualifications, historical cooperation performance, and material supply capacity of each supplier node are evaluated to generate a global matching score for that supplier node. Based on the global matching score of all supplier nodes, the global candidate node set is sorted and a supplier recommendation list is output.

8. A supply chain supplier precision matching system based on big data analysis, characterized in that, The system includes: The module is used to integrate procurement needs, supplier profiles, material master data, and historical transaction records to build a big data collection in the supply chain field. The processing module is used to construct a supply chain knowledge graph based on the big data set in the supply chain field, wherein the nodes include supplier entities, material entities, and purchaser entities, and the edges represent supply relationships, cooperation history, material matching relationships, and transaction evaluation data; The processing module is also used to identify and quantify the deviation risk level of the current procurement demand relative to the historical procurement benchmark pattern in the supply chain knowledge graph. The processing module is also used to calculate the supply network stability index and material substitution difficulty index of the materials in the current procurement list by mining the topology of the supply chain knowledge graph and analyzing the association rules of historical transaction data. The execution module is used to make judgments based on the deviation risk level, the supply network stability index, and the material substitution difficulty index, and dynamically select a supplier matching strategy: if the overall risk is controllable, the supply network is stable, and the material substitution difficulty is low, then global supplier matching is initiated based on multi-dimensional similarity matching of node attributes and relationships in the entire supply chain knowledge graph; if the overall risk level is high, the supply network is unstable, or the material substitution difficulty is high, then a trusted supplier priority recommendation mechanism based on historical cooperation subgraphs and performance big data is triggered.

9. A computer device comprising a memory and a processor, the memory storing code, characterized in that, The processor is configured to acquire the code and execute the supply chain supplier precision matching method based on big data analysis as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the supply chain supplier precision matching method based on big data analysis as described in any one of claims 1 to 7.

Images