Automobile industry supply chain information processing method based on knowledge graph

Abstract

The invention relates to the technical field of data processing, in particular to an automobile industry supply chain information processing method based on a knowledge graph, and the method comprises the steps: constructing the knowledge graph of an automobile industry supply chain; the static risk degree and the dynamic risk degree of any node in the knowledge graph are calculated, the static risk degree is used for reflecting the financial health degree, the technology maturity and the supply stability of the node, and the dynamic risk degree is in positive correlation with the edge weight between the node and the neighbor node and the static risk degree of the node. According to the invention, the historical comprehensive risk degree of the node is obtained by giving the weight ratio of the static risk degree and the dynamic risk degree of the node to the historical comprehensive risk degree, so that the risk label of the real-time comprehensive risk degree is judged, the accuracy of judging the real-time comprehensive risk degree is improved, and the reliability of the real-time comprehensive risk degree is improved. And the problem of risk identification errors caused by insufficient transparency among enterprises in automobile supply chain risk identification is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology. More specifically, the present invention relates to an automobile industry supply chain information processing method based on knowledge graph. Background Art

[0002] In the modern automotive industry, the complexity of the supply chain information system is increasing, involving multiple participants such as suppliers, manufacturers, distributors, logistics service providers and retailers. Each participant plays an indispensable role in the supply chain and is interconnected with each other. The stability of their operations directly affects the integrity and efficiency of the entire supply chain. However, with the acceleration of globalization, the vulnerability of the supply chain has also increased, and supply chain risk management has become one of the key areas for companies to maintain their competitiveness.

[0003] In the related technologies, for example, the Chinese patent application document with publication number CN109767213A discloses a supply chain information processing method, device, equipment and medium based on blockchain. The patent application document sets up smart contracts on the blockchain, manages the transaction process in the automotive industry supply chain through the smart contracts on the blockchain, and uses blockchain technology to ensure the authenticity and traceability of the entire process data, thereby improving the information transparency in the automotive industry supply chain and reducing the transaction risks in the automotive industry supply chain.

[0004] At present, the automotive industry's supply chain involves multiple links and a large number of suppliers. Lack of transparency between companies may cause potential risks to be ignored, which in turn leads to incorrect identification of supply chain risks, causing losses to companies. Therefore, how to improve the accuracy of supply chain risk identification has become an urgent problem to be solved in the automotive industry. Summary of the invention

[0005] The present invention provides an automobile industry supply chain information processing method based on knowledge graph, aiming to solve the problem in related technologies that the current automobile industry supply chain involves multiple links and a large number of suppliers, and insufficient transparency between enterprises may cause potential risks to be ignored, leading to errors in supply chain risk identification, thereby causing losses to the enterprise.

[0006] The present invention provides an automobile industry supply chain information processing method based on a knowledge graph, including: constructing a knowledge graph of the automobile industry supply chain; calculating the static risk level and dynamic risk level of any node in the knowledge graph, wherein the static risk level is used to reflect the financial health, technical maturity and supply stability of the node, and the dynamic risk level is positively correlated with the edge weight between the node and its neighboring nodes and the static risk level of the node, and the edge weight reflects the industrial association between the node and its neighboring nodes; setting corresponding weights based on the static risk level and dynamic risk level of the node, and calculating the node Historical comprehensive risk level , the calculation formula is: , , ; In the formula, For Node The historical comprehensive risk level of and Don't be a node The weight of the static risk level and the weight of the dynamic risk level, , is a natural constant, and Node The static risk level and dynamic risk level of the supply chain are calculated based on the real-time node data, and the risk label of the real-time node is determined based on the comparison between the real-time comprehensive risk level and the historical comprehensive risk level. This method comprehensively considers static risk and dynamic risk, where static risk reflects long-term stable factors such as the financial health, technological maturity, and supply stability of the node, while dynamic risk considers the edge weight (industry relevance) between the node and its neighboring nodes and the dynamic changes of the node itself. This comprehensive evaluation method can more accurately capture the impact of static factors and external dynamic changes on supply chain node risks, and help to more comprehensively understand the potential risks of nodes.

[0007] Furthermore, determining the risk label of the real-time node includes: if the real-time comprehensive risk level is greater than or equal to the historical comprehensive risk level, determining the risk label of the node to be 1 (risk has occurred); if the real-time comprehensive risk level is less than the historical comprehensive risk level, determining the risk label of the node to be 0 (risk has not occurred).

[0008] Furthermore, we determine the weight and weight , including: constructing the historical comprehensive risk level and maximizing the inter-class variance according to the static risk level and dynamic risk level of each node and the corresponding risk label, determining the optimal weight of the static risk level and dynamic risk level of each node for the calculation of the historical comprehensive risk level, the optimal weight includes the weight and weight .

[0009] Furthermore, the dynamic risk level of the node is calculated using the following formula: ; In the formula, For Node The degree of dynamic risk; For Node and its neighbor nodes The edge rights between For Node The static risk level of For Node The number of neighbor nodes; For Node The set of neighbor nodes (i.e., nodes directly connected nodes); is a hyperbolic tangent function. By maximizing the inter-class variance to adjust the weights, the weights can better reflect the actual situation, adapt to the impact of static and dynamic risks on decision-making in different scenarios, and can be adjusted according to historical data to optimize the risk management model.

[0010] Furthermore, the node and its neighbor nodes The edge weight The calculation formula is: ; In the formula, For Node and its neighbor nodes Average frequency of project collaboration between For Node and its neighbor nodes The number of cooperative projects between For Node and its neighbor nodes The amount of cooperative transactions between For Node The calculation of edge weight combines three factors: cooperation frequency, number of cooperation projects and cooperation amount, so that the calculated edge weight more comprehensively reflects the cooperation intensity and closeness between nodes.

[0011] Furthermore, computing nodes The static risk level is calculated as follows: ; In the formula, For Node The static risk level of For Node No. Features for nodes The weight of risk; For Node No. characteristics, The natural constant is the exponential function of the base, When represents financial health, technology maturity and supply stability respectively. This method combines the characteristics of three different dimensions, namely financial health, technology maturity and supply stability, and comprehensively considers multiple factors that affect node risk. Through this multi-dimensional consideration, the model can more comprehensively and accurately evaluate the static risk of the node.

[0012] Furthermore, the node No. Features for nodes The formula for calculating the risk weight is: ; In the formula, For Node No. Features for nodes The weight of risk, For Node No. Characteristics For Node No. The mutual information entropy between the features and the corresponding risk labels, where the risk label is 0, indicating that the risk has not occurred, and the risk label is 1, indicating that the risk has occurred. When , they represent financial health, technological maturity, and supply stability, respectively. Mutual information entropy measures the interdependence between features and risk labels. If a feature has a high mutual information entropy with a risk label, it means that the feature is more closely associated with the risk, that is, the feature contributes more to risk assessment. In this way, the weights can reflect the importance of each feature to risk prediction. Features with high mutual information entropy will be given higher weights, while features with less impact on risk prediction will be given lower weights.

[0013] Furthermore, constructing a knowledge graph of the automotive industry supply chain includes: collecting automotive industry supply chain data, wherein the supply chain data includes internal characteristics of the constituent entities of the supply chain, wherein the internal characteristics include debt-to-asset ratio, current ratio, quick ratio, R&D investment ratio, automation equipment ratio, on-time delivery rate, order change rate and inventory turnover rate, etc.; collecting industry association data between constituent entities, wherein the industry association data includes the average project cooperation frequency, the number of cooperation projects, and the amount of cooperation transaction volume; obtaining the constituent entities of the supply chain, wherein the constituent entities include suppliers, manufacturers, distributors, retailers, logistics service providers and consumers, and constructing a knowledge graph according to the industry association data between the constituent entities.

[0014] Furthermore, a knowledge graph is constructed according to the industrial correlation data between the constituent entities, including: taking each constituent entity as a node and the relationship between the constituent entities as an edge, where the relationship includes cooperative relationship, supply and demand relationship, and purchasing relationship, to construct a knowledge graph of the automotive industry supply chain.

[0015] Beneficial effects: By constructing a knowledge graph of the automotive industry supply chain and introducing a method to calculate the dynamic risk level and static risk level of nodes in the knowledge graph, and by assigning a weight ratio between the static risk level and dynamic risk level of the node for the calculation of the historical comprehensive risk level, the historical comprehensive risk level of the node is obtained to determine the risk label of the real-time comprehensive risk level, thereby improving the accuracy of the judgment of the real-time comprehensive risk level, reducing the risk identification errors caused by insufficient transparency between enterprises in the automotive supply chain risk identification, and improving the risk identification accuracy of the automotive industry supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The following detailed description is read with reference to the accompanying drawings, which illustrate several embodiments of the present invention in an exemplary and non-limiting manner, and in which like or corresponding reference numerals represent like or corresponding parts, wherein: Figure 1 is a flow chart schematically illustrating a method of determining a risk label of a real-time node according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0018] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0019] like Figure 1 As shown, S101: Constructing a knowledge graph for the automotive industry supply chain.

[0020] Specifically, the automotive industry supply chain data is collected, and the supply chain data includes the internal characteristics of the constituent entities of the supply chain, wherein the internal characteristics include debt-to-asset ratio, current ratio, quick ratio, R&D investment ratio, automation equipment ratio, delivery on-time rate, order change rate and inventory turnover rate, etc. The industry association data between the constituent entities is collected, wherein the industry association data includes the average frequency of project cooperation, the number of cooperative projects, and the amount of cooperative transaction volume. Then the constituent entities of the supply chain are obtained, and the constituent entities include suppliers, manufacturers, distributors, retailers, logistics service providers and consumers, and a knowledge graph is constructed according to the industry association data between the constituent entities. Specifically, each constituent entity is used as a node, and the relationship between each constituent entity is used as an edge, wherein the relationship includes cooperation relationship, supply and demand relationship, and purchase relationship, to construct a knowledge graph of the automotive industry supply chain.

[0021] S102: Calculate the static risk level of the nodes in the knowledge graph.

[0022] Specifically, the internal features of each node in the knowledge graph of the automotive industry supply chain obtained above are integrated to calculate the financial health, technology maturity and supply stability of each node. The mutual information entropy between the financial health, technology maturity and supply stability of the node and the corresponding risk label (label 0 means that the risk has not occurred, label 1 means that the risk has occurred) is calculated to calculate the static risk level of each node.

[0023] In one embodiment, taking any node in the knowledge graph as an example, the static risk level of the node is calculated, and the calculation formula is: ; In the formula, For Node The static risk level of For Node No. Features for nodes The weight of risk; For Node No. characteristics, The natural constant is the exponential function of the base, and represent financial health, technological maturity, and supply stability, respectively. For Node The sum of the weighted characteristics of the enterprise. The larger the value, the greater the sum of the enterprise's financial health, technological maturity and supply stability, and the smaller the static risk of the enterprise. The smaller the value, the smaller the sum of the enterprise's financial health, technological maturity and supply stability, and the greater the platform risk of the enterprise.

[0024] In one embodiment, the node No. Features for nodes The formula for calculating the risk weight is: ; In the formula, For Node No. Features for nodes The weight of risk, For Node No. Characteristics For Node No. The mutual information entropy between the features and the corresponding risk labels, where the risk label is 0, indicating that the risk has not occurred, and the risk label is 1, indicating that the risk has occurred. , , and represent financial health, technological maturity, and supply stability, respectively. The larger the value, the node No. The greater the correlation between a feature and risk, the greater the weight assigned. The smaller the value, the node No. The less the correlation between a characteristic and risk, the smaller the weight assigned to it.

[0025] In one embodiment, according to the node The financial health is calculated by using the debt-to-asset ratio, current ratio, and quick ratio. The calculation formula is: ; In the formula, For Node Financial health of For Node Current ratio; For Node Quick ratio; For Node debt-to-asset ratio; is the normalization function.

[0026] in, The larger the value, the node The stronger the short-term debt repayment ability, the greater the financial health. The smaller the value, the node The weaker the short-term debt repayment ability, the lower the financial health; The larger the value, the node The stronger the short-term debt repayment ability, the greater the financial health. The smaller the value, the smaller the node The weaker the short-term debt repayment ability, the lower the financial health; The larger the value, the node The higher the debt level, the lower the financial health. The smaller the value, the node The lower the debt level, the greater the financial health.

[0027] In one embodiment, according to the node The technology maturity is calculated based on the proportion of R&D investment and the proportion of automation equipment. The calculation formula is: ; In the formula, For Node The maturity of technology; For Node R&D investment share; For Node The proportion of automated equipment; is the normalization function.

[0028] in, The larger the value, the higher the enterprise's investment in technological innovation and the greater the technological maturity. The smaller the value, the lower the enterprise's investment in technological innovation and the lower the technological maturity; The larger the value is, the higher the level of automated production and the greater the technological maturity of the enterprise. The smaller the value, the lower the level of automated production and the lower the technological maturity of the enterprise.

[0029] In one embodiment, according to the node The supply stability is calculated based on the on-time delivery rate, order change rate and inventory turnover rate. The calculation formula is: ; In the formula, For Node Supply stability; For Node On-time delivery rate; For Node inventory turnover rate; is the order change rate; is the normalization function.

[0030] in, For Node The on-time delivery rate can be obtained by the ratio of the number of on-time delivery orders to the total number of orders. The larger the value, the stronger the company's ability to deliver on time and the greater the supply stability. The smaller the value, the weaker the enterprise's ability to deliver on time and the smaller the supply stability; For Node The inventory turnover rate can be obtained by dividing the cost of sales by the average inventory. The larger the value, the faster the company's inventory turnover and the greater the supply stability. The smaller the value, the slower the inventory turnover of the enterprise and the smaller the supply stability; is the order change rate, which can be obtained by the ratio of the standard deviation of the order quantity to the mean of the order quantity. The larger the value, the greater the order fluctuation of the enterprise and the smaller the supply stability. The smaller the value of, the smaller the order fluctuation of the enterprise is, and the greater the supply stability is. So far, the static risk level of each node in the knowledge graph of the automotive industry supply chain can be obtained.

[0031] S103: Calculate the dynamic risk level of the node.

[0032] Specifically, firstly, according to the nodes in the automotive industry knowledge graph The industry association data between its neighbor nodes (average frequency of project cooperation, number of cooperative projects, and cooperative transaction amount) are used to calculate the edge weight between nodes. The calculation formula is: ; In the formula, For Node and its neighbor nodes The edge right between For Node and its neighbor nodes Average frequency of project collaboration between For Node and its neighbor nodes The number of cooperative projects between For Node and its neighbor nodes The amount of cooperative transactions between For Node The set of neighbor nodes.

[0033] in, The larger the value, the node and its neighbor nodes The greater the degree of industrial correlation between nodes, the greater the edge weight between nodes. The smaller the value, the node and its neighbor nodes The smaller the degree of industry correlation, the smaller the edge weight between nodes; The larger the value, the node and its neighbor nodes The greater the degree of industrial correlation between nodes, the greater the edge weight between nodes. The smaller the value, the node and its neighbor nodes The greater the degree of industry correlation, the smaller the edge weight between nodes; The larger the value, the node and its neighbor nodes The greater the degree of industrial correlation between nodes, the greater the edge weight between nodes. The smaller the value, the node and its neighbor nodes The greater the degree of industrial correlation, the smaller the edge weight between nodes.

[0034] In one embodiment, the dynamic risk level of the node is calculated based on the number of neighbor nodes of the node obtained in the knowledge graph, the static risk level of the node, and the edge weight between the node and its neighbor nodes obtained above. The calculation formula is: ; In the formula, For Node The degree of dynamic risk; For Node and its neighbor nodes The edge rights between For Node The static risk level of For Node The number of neighbor nodes; For Node The set of neighbor nodes (i.e., nodes directly connected nodes); is the hyperbolic tangent function. Among them, Neighbor node Propagate to nodes The risk level of the neighbor node is higher. Propagate to nodes The greater the risk level, the smaller the value, indicating that the neighbor node Propagate to nodes The lower the risk level, the lower the risk level. So far, the dynamic risk level of each node in the knowledge graph of the automotive industry supply chain can be obtained.

[0035] S104: Calculate the historical comprehensive risk level of the node.

[0036] In one embodiment, the corresponding weights are set based on the static risk level and dynamic risk level of the node, and the node is calculated. Historical comprehensive risk level , the calculation formula is: , , ; In the formula, For Node The historical comprehensive risk level of and Don't be a node The weight of the static risk level and the weight of the dynamic risk level, , is a natural constant, and Node The static risk level and dynamic risk level of the company.

[0037] In one embodiment, the weight and weight The determination method is as follows: static risk level changes less in the short term and often has a greater impact on corporate decisions and results, while dynamic risk level fluctuates with changes in the external environment and its impact is less predictable in the short term. Therefore, the dynamic risk level has a smaller weight than the static risk level. Therefore, the weight can be set is 0.7, weight The value of is 0.3.

[0038] In another embodiment, the weight setting method is based on past experience, and the weight is fixed, which may not adapt to the complexity of specific scenarios and lacks flexibility and adjustability. Therefore, another weight setting method is provided. and weight The determination method is as follows: construct the historical comprehensive risk level according to the static risk level and dynamic risk level of each node and the corresponding risk label and maximize the inter-class variance, determine the optimal weight of the static risk level and dynamic risk level of each node for the calculation of the historical comprehensive risk level, and the optimal weight includes weight and weight . By maximizing the inter-class variance to adjust the weights, the weights can better reflect the actual situation and adapt to the impact of static and dynamic risks on decision-making in different scenarios. They can also be adjusted based on historical data to optimize the risk management model. The historical comprehensive risk level calculated using the optimal weight is used to determine the risk label of the real-time comprehensive risk level, which improves the accuracy of the comprehensive risk level judgment and reduces the risk identification errors caused by insufficient transparency between enterprises in the automotive supply chain risk identification, thereby improving the risk identification accuracy of the automotive industry supply chain.

[0039] In one embodiment, the maximum inter-class variance of class 0 (risk has not occurred) and class 1 (risk has occurred) in the risk label of the node is calculated, including: firstly, the static risk degree and dynamic risk degree of each node in the automotive industry knowledge graph after each historical data update and the corresponding risk label are counted to obtain the risk label histogram of each node, and the probability of different risk labels in each node in the histogram is calculated, and the probability of different risk labels in each node is calculated. For example, the risk label of node The calculation formulas for the probabilities of medium risk label 0 and risk label 1 in the risk histogram are: ; In the formula, and Node The probability of medium risk label 0 and risk label 1 in the risk histogram; and Node The number of medium risk label 0 and risk label 1 in the risk histogram; For Node The total number of risk labels in the medium. Then calculate the historical comprehensive risk level mean of class 0, the historical comprehensive risk level mean of class 1, and the historical comprehensive risk level mean of all risk labels, and finally calculate the inter-class variance.

[0040] In one embodiment, the calculation formula of the between-class variance is: In the formula, For Node The inter-class variance of the risk labels in class 0 (risk did not occur) and class 1 (risk occurred), For Node The historical comprehensive risk level average of all risk labels, For Node The historical comprehensive risk level average of the risk label of 0, For Node The historical comprehensive risk level average of the risk label of 1, and Node The probability of risk label 0 and risk label 1 in the risk histogram.

[0041] In one embodiment, the calculation formula for the average value of the historical comprehensive risk level of all risk tags is: ; In the formula, For Node The historical comprehensive risk level average of all risk labels, For Node The total number of medium risk labels, For Node No. The historical comprehensive risk level after the data update.

[0042] In one embodiment, the calculation formula of the historical comprehensive risk level average of category 0 is: , where For Node The historical comprehensive risk level average of the risk label of 0, For Node Belongs to the sample set with risk label 0, For Node The total number of medium risk labels is 0, For Node No. The historical comprehensive risk level after the data update.

[0043] In one embodiment, the calculation formula of the historical comprehensive risk level average of category 1 is: , where For Node The historical comprehensive risk level average of the risk label of 1, For Node The sample set with risk label 1, For Node No. The historical comprehensive risk level after the data update, For Node The total number of medium risk labels is 1. At this point, the historical comprehensive risk level can be calculated using the optimal weight obtained by maximizing the inter-class variance.

[0044] S105: Determine the risk label of the real-time node.

[0045] In one embodiment, the real-time comprehensive risk level is calculated based on the real-time node data, and the risk label of the real-time node is determined based on the comparison result between the real-time comprehensive risk level and the historical comprehensive risk level. Specifically, if the real-time comprehensive risk level is greater than or equal to the historical comprehensive risk level, the risk label of the node is determined to be 1 (risk has occurred); if the real-time comprehensive risk level is less than the historical comprehensive risk level, the risk label of the node is determined to be 0 (risk has not occurred).

[0046] By adopting the above steps and constructing a knowledge graph of the automotive industry supply chain, the entities in the supply chain and the relationships between entities are structured, which improves the transparency between entities in the automotive industry supply chain. The calculation method of the dynamic risk level and static risk level of the nodes in the knowledge graph is introduced. The historical comprehensive risk level of the node is obtained by assigning a weight ratio of the static risk level and dynamic risk level of the node to the calculation of the historical comprehensive risk level, so as to judge the risk label of the real-time comprehensive risk level, improve the accuracy of the judgment of the real-time comprehensive risk level, reduce the risk identification errors caused by insufficient transparency between enterprises in the automotive supply chain risk identification, and improve the accuracy of risk identification in the automotive industry supply chain.

[0047] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0048] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0049] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent application. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for processing information in the automotive industry supply chain based on knowledge graph, characterized in that: include: Build a knowledge graph for the automotive industry supply chain; Calculate the static risk level and dynamic risk level of any node in the knowledge graph, where the static risk level is used to reflect the financial health, technical maturity and supply stability of the node, and the dynamic risk level is positively correlated with the edge weight between the node and its neighboring nodes and the static risk level of the node, and the edge weight reflects the industrial correlation between the node and its neighboring nodes; Based on the static risk level and dynamic risk level of the node, the corresponding weights are set to calculate the historical comprehensive risk level of the node. The calculation formula is: , , ; In the formula, For Node The historical comprehensive risk level of and Don't be a node The weight of the static risk level and the weight of the dynamic risk level, , is a natural constant, and Node The static risk level and dynamic risk level of the company; The real-time comprehensive risk level is calculated based on the real-time node data, and the risk label of the real-time node is determined based on the comparison between the real-time comprehensive risk level and the historical comprehensive risk level.

2. The automotive industry supply chain information processing method based on knowledge graph according to claim 1 is characterized in that: Determine the risk label of the live node, including: If the real-time comprehensive risk level is greater than or equal to the historical comprehensive risk level, the risk label of the node is determined to be 1 (risk has occurred); If the real-time comprehensive risk level is less than the historical comprehensive risk level, the risk label of the node is determined to be 0 (no risk has occurred).

3. The automotive industry supply chain information processing method based on knowledge graph according to claim 1 is characterized in that: Determine the weight and weight ,include: According to the static risk level and dynamic risk level of each node and the corresponding risk label, the historical comprehensive risk level is constructed and the inter-class variance is maximized, and the optimal weights of the static risk level and dynamic risk level of each node for the calculation of the historical comprehensive risk level are determined. The optimal weights include weight and weight .

4. The automotive industry supply chain information processing method based on knowledge graph according to claim 1 is characterized in that: Calculate the dynamic risk level of the node, the calculation formula is: ； In the formula, For Node The degree of dynamic risk; For Node and its neighbor nodes The edge rights between For Node The static risk level of For Node The number of neighbor nodes; For Node The set of neighbor nodes (i.e., nodes directly connected nodes); is the hyperbolic tangent function.

5. The automotive industry supply chain information processing method based on knowledge graph according to claim 4 is characterized in that: node and its neighbor nodes The edge weight The calculation formula is: ； In the formula, For Node and its neighbor nodes Average frequency of project collaboration between For Node and its neighbor nodes The number of cooperative projects between For Node and its neighbor nodes The amount of cooperative transactions between For Node The set of neighbor nodes.

6. The automotive industry supply chain information processing method based on knowledge graph according to claim 1 is characterized in that: Compute Node The static risk level is calculated as follows: ； In the formula, For Node The static risk level of For Node No. Features for nodes The weight of risk; For Node No. characteristics, The natural constant is the exponential function of the base, , , and represent financial health, technological maturity, and supply stability, respectively.

7. The automotive industry supply chain information processing method based on knowledge graph according to claim 6 is characterized in that: node No. Features for nodes The formula for calculating the risk weight is: ； In the formula, For Node No. Features for nodes The weight of risk, For Node No. Characteristics For Node No. The mutual information entropy between the features and the corresponding risk labels, where the risk label is 0, indicating that the risk has not occurred, and the risk label is 1, indicating that the risk has occurred. , , and represent financial health, technological maturity, and supply stability, respectively.

8. The automotive industry supply chain information processing method based on knowledge graph according to claim 1 is characterized in that: Build a knowledge graph of the automotive industry supply chain, including: Collecting automotive industry supply chain data, wherein the supply chain data includes internal characteristics of the supply chain components, wherein the internal characteristics include debt-to-asset ratio, current ratio, quick ratio, R&D investment ratio, automation equipment ratio, on-time delivery rate, order change rate, and inventory turnover rate; Collect industry-related data between constituent entities, including average project cooperation frequency, number of cooperation projects, and cooperation transaction volume and amount; Obtain the constituent entities of the supply chain, including suppliers, manufacturers, distributors, retailers, logistics service providers and consumers, and construct a knowledge graph based on the industrial association data between the constituent entities.

9. The automotive industry supply chain information processing method based on knowledge graph according to claim 8 is characterized in that: Construct a knowledge graph based on the industry-related data between constituent entities, including: With each component entity as a node and the relationship between the components as an edge, including cooperative relationships, supply and demand relationships, and purchasing relationships, a knowledge graph of the automotive industry supply chain is constructed.

Images