Industrial cluster industrial chain construction method, apparatus and device, and medium

Abstract

The invention relates to the technical field of industrial clusters, and discloses an industrial cluster industrial chain construction method and device, equipment and a medium, and the method comprises the steps: obtaining supply associated data and property right associated data of an enterprise of an industry category to which a target industrial cluster belongs; constructing a directed contact network based on the supply association data; constructing an undirected contact network based on the property right associated data; calculating property right connection degrees among the nodes in the undirected connection network, and accumulating the property right connection degrees to paths with corresponding node relations in the directed connection network according to a preset weight to generate a multi-source data enhanced network; and traversing all node paths in the multi-source data enhanced network, calculating the accumulated connection strength of each node path, and determining the path with the highest accumulated connection strength as the industrial cluster core industrial chain. According to the method, the universality, professional interpretability and objective accuracy of the constructed industrial chain are improved.

Description

Technical Field

[0001] This invention relates to the field of industrial cluster technology, and more specifically, to a method, apparatus, equipment and medium for constructing an industrial cluster supply chain. Background Technology

[0002] Industrial clusters are a crucial pillar of regional economic development, characterized by their wide scope, diverse product types (e.g., high-end equipment manufacturing clusters themselves encompass multiple product lines and may have multiple industrial chains), and complex regional variations. Therefore, quantitative analysis of industrial clusters in different regions is particularly important, and extracting key industrial chains from the network of industrial cluster connections can provide a crucial tool for industrial cluster analysis.

[0003] Currently, the main technical approaches to building industrial chains include: relying on deep learning algorithms to achieve entity recognition and relationship extraction to construct knowledge graphs; aligning entities by identifying keywords through predefined industrial chain graph structures; and directly predefining entity-relationship graphs and incorporating quantitative indicators. However, the relationships between entities obtained by deep learning-based methods are not clearly quantified, lacking obvious upstream and downstream relationships, making it difficult to provide professional knowledge explanations in industrial economic applications. Methods based on predefined structures suffer from strong subjectivity, high requirements for prior knowledge, and questionable authenticity of the identified entity relationships. In summary, in constructing industrial cluster industrial chains that truly reflect industrial relationships and possess high interpretability, existing technologies suffer from drawbacks such as strong subjectivity of predefined structures, insufficient universality, and low professional interpretability due to pure algorithmic reliance. Summary of the Invention

[0004] To overcome the shortcomings of existing industry chain construction technologies, such as strong subjectivity, low universality, and low professional interpretability, this invention proposes the following technical solution: Firstly, this invention proposes a method for constructing an industrial cluster supply chain, comprising: Obtain supply-related data and ownership-related data of enterprises in the industry category to which the target industrial cluster belongs; A directed connection network is constructed based on the aforementioned supply-related data; Construct an undirected network based on the aforementioned property rights association data; Calculate the property rights connection degree between each node in the undirected network, and accumulate the property rights connection degree to the path with corresponding node relationship in the directed network according to the preset weight to generate a multi-source data reinforcement network; Traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

[0005] As a preferred technical solution, the acquisition of supply-related data and property rights-related data of enterprises in the industry category to which the target industrial cluster belongs includes: Obtain descriptive text related to the target industry cluster, and extract keyword datasets from the descriptive text using a pre-defined large language model; The keyword dataset is matched with the preset national economic industry classification codes to determine the set of industry categories involved in the target industrial cluster; Obtain a list of surviving companies within the target region based on the aforementioned set of industry categories; The information on the related initiating companies, related receiving companies, and contact frequency of the companies in the list of existing companies is obtained as the property rights association data, and the information on the purchasing companies, the purchasing companies, and the purchasing frequency is obtained as the supply association data.

[0006] As a preferred technical solution, a directed connection network is constructed based on the supply-related data, including: Quantitative processing is performed on supply-related data, including information on purchasing companies, information on companies being purchased, and purchase frequency, to construct a procurement relationship matrix; Based on the preset industry classification standards, the purchasing enterprise information and purchased enterprise information in the purchasing connection matrix are classified, and the purchasing frequency of records belonging to the same purchasing industry classification and the same purchased industry classification is accumulated to generate a directed connection network.

[0007] As a preferred technical solution, after generating the directed connection network, the method further includes: Traverse the directed relationship network. If a record has the same industry classification for purchasing as the industry classification for the purchased item, remove that record from the directed relationship network. Check if there are bidirectional connection records in the directed connection network that represent the purchasing industry category and the purchased industry category, respectively. If so, calculate the difference in purchasing frequency after summing the two directions, retain the record with the larger summed purchasing frequency, and update the purchasing connection degree of that record to the difference. If the summed purchasing frequencies of the two directions are the same, remove the bidirectional connection record to obtain the optimized directed connection network.

[0008] As a preferred technical solution, an undirected network is constructed based on the aforementioned property rights association data, including: The property rights association data, including information on related initiating companies, related receiving companies, and contact frequency, are quantified to construct a property rights association matrix. Based on the preset industry classification standards, the information of the associated initiating enterprises and the information of the associated receiving enterprises in the property rights connection matrix are mapped to the corresponding industry classifications, thereby generating an undirected connection network.

[0009] As a preferred technical solution, the property rights connection degree between each node in the undirected network is calculated, and the property rights connection degree is accumulated according to a preset weight onto the paths with corresponding node relationships in the directed network to generate a multi-source data reinforcement network, including: Traverse each directed path in the directed network and identify the starting and ending nodes corresponding to the directed path; Using the starting node and the ending node as search criteria, retrieve the corresponding connection records in the property rights connection matrix; If a corresponding contact record is found, the contact frequency in the contact record is extracted as the property rights connection strength. The property rights connection strength is multiplied by a preset strengthening weight coefficient to obtain the strengthening increment value. The reinforcement increment is accumulated onto the corresponding directed path, and the edge values ​​of the corresponding directed path are updated until all directed paths have been traversed, thus obtaining the multi-source data reinforcement network.

[0010] As a preferred technical solution, all node paths in the multi-source data reinforcement network are traversed, the cumulative connection strength of each node path is calculated, and the path with the highest cumulative connection strength is identified as the core industrial chain of the industrial cluster, including: Based on the multi-source data enhancement network, a node connectivity matrix is ​​constructed; the numerical values ​​of the elements in the node connectivity matrix represent the connection strength between the corresponding nodes. The values ​​of the elements in the node connectivity matrix are inverted to construct a path calculation matrix for path search. All nodes in the multi-source data augmentation network are traversed sequentially as intermediate nodes. For any pair of starting and ending nodes, the sum of the path value from the starting node to the intermediate node and the path value from the intermediate node to the ending node in the path calculation matrix is ​​calculated. Determine whether the sum of the path values ​​is less than the current path value from the starting node to the ending node recorded in the path calculation matrix. If so, update the current path value with the sum of the path values ​​until all nodes have been traversed. After traversing the path, obtain the absolute value of the path value between each pair of starting and ending nodes in the path calculation matrix. Determine the absolute value as the cumulative connection strength of the corresponding node path, and determine the node path with the largest cumulative connection strength as the core industrial chain of the industrial cluster.

[0011] Secondly, the present invention also proposes an industrial cluster supply chain construction device, applied in the industrial cluster supply chain construction method as described in any of the embodiments of the first aspect, comprising: The acquisition module is used to acquire supply-related data and property rights-related data of enterprises in the industry category to which the target industrial cluster belongs; The first construction module is used to construct a directed connection network based on the supply-related data; The second construction module is used to construct an undirected network based on the property rights association data; The first calculation module is used to calculate the property rights connection degree between each node in the undirected network, and to accumulate the property rights connection degree to the path with corresponding node relationship in the directed network according to the preset weight, so as to generate a multi-source data reinforcement network. The second calculation module is used to traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

[0012] Thirdly, the present invention also proposes an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the operations performed by the industrial cluster supply chain construction method as described in any of the embodiments of the first aspect.

[0013] Fourthly, the present invention also proposes a computer-readable storage medium storing a program, the program being executed by a processor as performed by the industrial cluster supply chain construction method described in any of the embodiments of the first aspect.

[0014] The beneficial effects of the present invention include at least the following: This invention acquires supply-related data and property rights-related data to construct directed and undirected connection networks, respectively. It then uses the calculated undirected property rights connection degree to weight and accumulate the directed supply network, generating a multi-source data-enhanced network. This network leverages objectively existing property rights and capital connections to compensate for the shortcomings of single transaction data in describing industrial relationships, effectively constructing a physical network that reflects the real technological and economic connections between industries. This avoids the subjectivity arising from predefined industrial chain structures in traditional methods. Furthermore, by traversing the enhanced network and calculating the cumulative connection strength, the core industrial chain is determined. This achieves quantitative data-driven development instead of pure algorithmic deduction or manual pre-setting, ensuring that the constructed industrial chain is entirely based on the strength of actual industrial connection data. This eliminates reliance on prior templates for specific industries, significantly improving the universality of the constructed industrial chain while ensuring its high degree of professional interpretability and objective accuracy. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the industrial cluster supply chain construction method provided in this embodiment of the invention.

[0016] Figure 2This is a schematic diagram of the updated procurement category contact network provided in an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of the core industry path in the directed network of the automotive industry cluster provided in an embodiment of the present invention.

[0018] Figure 4 This is an architectural diagram of the industrial cluster supply chain construction device provided in an embodiment of the present invention.

[0019] Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred technical solutions. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred technical solutions are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0021] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0022] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0023] Example 1 This embodiment proposes a method for constructing an industrial cluster supply chain, such as... Figure 1 As shown, Figure 1 This is a flowchart illustrating a method for constructing an industrial cluster supply chain provided in this embodiment. The method includes the following steps: S1: Obtain supply-related data and ownership-related data of enterprises in the industry category to which the target industrial cluster belongs; S2: Construct a directed connection network based on the aforementioned supply-related data; S3: Construct an undirected network based on the aforementioned property rights association data; S4: Calculate the property rights connection degree between each node in the undirected connection network, and accumulate the property rights connection degree to the path with corresponding node relationship in the directed connection network according to the preset weight to generate a multi-source data reinforcement network; S5: Traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

[0024] Understandably, by acquiring supply-related data and property rights-related data, directed and undirected connection networks are constructed respectively. The calculated undirected property rights connection degree is then used to weight and accumulate the directed supply network, generating a multi-source data-reinforced network. This leverages objectively existing property rights and capital connections to compensate for the shortcomings of single transaction data in describing industrial relationships, effectively constructing a physical network that reflects the real technological and economic connections between industries. This avoids the subjectivity arising from predefined industrial chain structures in traditional methods. Furthermore, by traversing the reinforced network and calculating the cumulative connection strength to determine the core industrial chain, quantitative data-driven approaches replace pure algorithmic deduction or manual pre-setting. This ensures that the constructed industrial chain is entirely based on the strength of actual industrial connection data, eliminating reliance on specific industry-specific prior templates. While significantly improving the universality of the constructed industrial chain, it also ensures high professional interpretability and objective accuracy.

[0025] Example 2 This embodiment is an improvement on the industrial cluster supply chain construction method proposed in Embodiment 1.

[0026] In this embodiment, obtaining the supply-related data and property rights-related data of enterprises in the industry category to which the target industrial cluster belongs includes: Obtain descriptive text related to the target industry cluster, and extract keyword datasets from the descriptive text using a pre-defined large language model; The keyword dataset is matched with the preset national economic industry classification codes to determine the set of industry categories involved in the target industrial cluster; Obtain a list of surviving companies within the target region based on the aforementioned set of industry categories; The information on the related initiating companies, related receiving companies, and contact frequency of the companies in the list of existing companies is obtained as the property rights association data, and the information on the purchasing companies, the purchasing companies, and the purchasing frequency is obtained as the supply association data.

[0027] As an example, in the specific implementation process, the following steps are taken: First, descriptive texts related to the industrial cluster (such as policies, technical standards, etc.) are collected. Keyword datasets are extracted using a large language model and matched with the national economic industry classification codes to obtain the national standard category G involved in the industrial cluster. For the research area, a list of all existing enterprises E is crawled through the national standard category G. Then, the industrial headquarters branch contact data table ZF, the industrial investment contact data table TZ, and the industrial procurement contact data table CG are constructed respectively.

[0028] Where ZF = {zbqy, fzqy, lx_zf, gygx_zf, gyfx_zf, gbfl_zb, gbfl_fz}, zbqy is the headquarters enterprise, fzqy is the branch enterprise, the total-to-branch connection degree lx_zf∈(0,1), the headquarters-branch supply relationship gygx_zf takes a value of 0.5 to express a fuzzy supply relationship, and the headquarters-branch supply direction gyfx_zf takes a value of 0 to express an unclear supply direction; TZ = {tzqy, btzqy, lx_tz, gygx_tz, gyfx_tz, gbfl_tz, gbfl_btz}, where tzqy is the investing company, btzqy is the invested company, the investment correlation degree lx_tz ∈ (0, 1), the supply relationship gygx_tz takes a value of 0.5, and the supply direction gyfx_tz takes a value of 0; CG = {cgqy, bcgqy, lx_cg, gygx_cg, gyfx_cg, gbfl_cg, gbfl_bcg}, where cgqy is the purchasing company, bcgqy is the company being purchased from, lx_cg is the frequency of actual purchases, gygx_cg is set to 1 to indicate a clear supply relationship, and gyfx_cg is set to 1 to indicate a clear one-way supply direction.

[0029] Furthermore, to visually demonstrate the data structure, the following are examples of basic data table formats: An example of the Industry Headquarters-Branch Contact Data Table (ZF) is shown in Table 1 below: Table 1. Data on the Connections Between Industry Headquarters and Branches

[0030] An example of the Industrial Investment Linkage Data Sheet (TZ) is shown in Table 2 below: Table 2. Industry Investment Linkage Data Table

[0031] An example of an Industry Procurement Contact Data Sheet (CG) is shown in Table 3 below: Table 3 Industry Procurement Contact Data Table

[0032] It is understandable that this embodiment relies on dynamically changing enterprise entity data and has no predetermined relationship structure. Therefore, it can dynamically acquire the industrial chain at different times and in different regions, which has high flexibility. At the same time, by distinguishing between explicit procurement and supply relationships and vague property rights and supply relationships, it quantifies and constructs a network of real-existing industrial link entities, reducing the subjectivity brought about by predefined industrial chain structures. This provides a solid quantitative data foundation for the subsequent construction of core industrial chains of industrial clusters with high professional interpretability.

[0033] In this embodiment, a directed connection network is constructed based on the supply association data, including: Quantitative processing is performed on supply-related data, including information on purchasing companies, information on companies being purchased, and purchase frequency, to construct a procurement relationship matrix; Based on the preset industry classification standards, the purchasing enterprise information and purchased enterprise information in the purchasing connection matrix are classified, and the purchasing frequency of records belonging to the same purchasing industry classification and the same purchased industry classification is accumulated to generate a directed connection network.

[0034] In this embodiment, after generating the directed connection network, the method further includes: Traverse the directed relationship network. If a record has the same industry classification for purchasing as the industry classification for the purchased item, remove that record from the directed relationship network. Check if there are bidirectional connection records in the directed connection network that represent the purchasing industry category and the purchased industry category, respectively. If so, calculate the difference in purchasing frequency after summing the two directions, retain the record with the larger summed purchasing frequency, and update the purchasing connection degree of that record to the difference. If the summed purchasing frequencies of the two directions are the same, remove the bidirectional connection record to obtain the optimized directed connection network.

[0035] As an example, in the specific implementation process, a procurement contact database matrix CG containing m rows of information is first constructed, denoted as CG={cgqy, bcgqy, lx_cg, gygx_cg, gyfx_cg, gbfl_cg, gbfl_bcg}, where lx_cg represents the procurement contact degree, and gbfl_cg and gbfl_bcg are the national standard classifications of the purchasing and purchased enterprises, respectively. Then, based on the research classification granularity, the enterprises are categorized upwards to obtain the directed contact network C. G_LB, when any row i, j in matrix CG satisfies gbfl_cg(i)=gbfl_cg(j) and gbfl_bcg(i)=gbfl_bcg(j), the degree of connection after classification is calculated as lx_cg_lb=lx_cg(i)+lx_cg(j); at the same time, the classified supply relationship gygx_cg_lb=1 is defined to express an explicit supply relationship; the classified supply direction gyfx_cg_lb=1 is defined to express an explicit unidirectional supply direction. Then, special cases are handled as follows: if gbfl_cg_lb(i) = gbfl_bcg_lb(i), that is, when the purchasing and the purchased category are the same, the record is deleted; if there is a bidirectional relationship between the purchasing and purchased categories, that is, gbfl_cg_lb(i) = gbfl_bcg_lb(j) and gbfl_bcg_lb(i) = gbfl_cg_lb(j), the degree of relationship between the two is compared. If lx_cg_lb(i) > lx_cg_lb(j), then lx_cg_lb = lx_cg_lb(i) - lx_cg_lb(j) is updated and the direction is retained. If the two are equal, they are deleted at the same time. In this way, a directed relationship network with fixed direction is constructed.

[0036] To illustrate the matrix processing process more clearly, the following example data is used: Assume that the preprocessed procurement contact database CG contains the following data (the last two columns are the categorized industry categories a, b, c, etc.): CG= [Company Name 1, Company Name 2, 1, 1, 1, a, a] Company Name 1, Company Name 3, 1, 1, 1, a, b Company Name 1, Company Name 4, 1, 1, 1, a, b Company Name 2, Company Name 5, 4, 1, 1, a, c Company Name 2, Company Name 6, 1, 1, 1, a, c Company Name 6, Company Name 8, 2, 1, 1, c, d Company Name 8, Company Name 10, 1, 1, 1, d, e Company Name 9, Company Name 6, 3, 1, 1, d, c] By performing upward categorization and accumulation, we obtain CG_LB: For example, if there are two records between category a and category b (company 1->3, company 1->4), the cumulative connection degree is 1+1=2; the connection degree between category d and category c (company 9->6) is 3. The calculated CG_LB is: [[a, a, 1], [a, b, 2], [a, c, 5], [c, d, 2], [d, e, 1], [d, c, 3]] Records [a, a, 1] containing the same category of purchase and purchased item are deleted.

[0037] There are bidirectional records [c, d, 2] and [d, c, 3]. Comparison shows that 3 > 2, so the direction d->c is retained, and the connection degree is updated to 3-2=1. The final directed connection network matrix is: CG_LB_Final=[[a, b, 2], [a, c, 5], [d, e, 1], [d, c, 1]].

[0038] Understandably, by eliminating self-loop connections within the same category, the problem of not being able to effectively express procurement relationships at a specific classification granularity is solved; by calculating the difference to process bidirectional connections and retaining the direction with stronger connections, the problem of not being able to express clear unidirectional supply relationships in bidirectional connections is solved. Thus, a stable directed connection network that can objectively reflect the supply relationships of upstream and downstream industries and has eliminated self-loops and bidirectional offsetting interference is constructed, providing a data foundation with clear direction for subsequent industrial chain analysis.

[0039] In this embodiment, an undirected network is constructed based on the property rights association data, including: The property rights association data, including information on related initiating companies, related receiving companies, and contact frequency, are quantified to construct a property rights association matrix. Based on the preset industry classification standards, the information of the associated initiating enterprises and the information of the associated receiving enterprises in the property rights connection matrix are mapped to the corresponding industry classifications, thereby generating an undirected connection network.

[0040] As an example, in the specific implementation process, a headquarters-branch connection database ZF and an investment connection database TZ are first constructed. Taking the headquarters-branch connection as an example, an m-row, 7-column matrix ZF={zbqy, fzqy, lx_zf, gygx_zf, gyfx_zf, gbfl_zb, gbfl_fz} is obtained through digital quantification. Here, lx_zf represents the total connection degree, with 0 for no connection and 1 for a connection; gygx_zf represents the supply relationship, with a value of 0.5 to express a vague supply relationship; and gyfx_zf represents the supply direction, with a value of 0 to express an unclear supply direction. Similarly, the investment connection matrix TZ is constructed. When constructing the undirected network, all connections are treated as undirected edges. That is, in subsequent classification and matching calculations, the active and passive relationships between headquarters enterprises and branch enterprises (or investing enterprises and invested enterprises) are not distinguished; only the existence of a connection between two industry nodes is considered. Next, based on the research classification granularity, the enterprises are categorized upwards to obtain the enterprise category connection matrix (such as ZF_LB). When any row i, j in matrix ZF satisfies gbfl_zf(i)=gbfl_zf(j) and gbfl_bzf(i)=gbfl_bzf(j), the connection degree after categorization is calculated as lx_zf_lb=lx_zf(i)+lx_zf(j). The TZ matrix is ​​processed in this way to generate an undirected connection network containing fuzzy supply relationships.

[0041] The construction of an undirected network is illustrated using specific case data: Assume the following data is used to connect the headquarters and branches with the ZF database: ZF= [Company Name 1, Company Name 2, 1, 0.5, 0, a, a] Company Name 1, Company Name 3, 1, 0.5, 0, a, b Company Name 1, Company Name 4, 1, 0.5, 0, a, b Company Name 8, Company Name 10, 1, 0.5, 0, d, e Company Name 9, Company Name 6, 1, 0.5, 0, d, c] After upward classification and accumulation, we obtain ZF_LB: ZF_LB=[[a, a, 1], [a, b, 2], [d, e, 1], [d, c, 1]]. This matrix serves as the basis for subsequent reinforcement, where the connection degree (e.g., 2, 1) will be used as the base for weight calculation.

[0042] Understandably, in the industrial sector, headquarters branches and investment connections only demonstrate ownership ties between industries and do not directly indicate whether there are clear upstream and downstream relationships. However, according to the logic of industrial cluster development, companies typically expand into related upstream and downstream industries during their expansion, resulting in ambiguous supply relationships. Therefore, treating these two types of connection data as an undirected network and assigning them specific fuzzy supply relationship values ​​can uncover potential industrial linkages. Stable ownership data can compensate for the shortcomings of single transaction data, providing reliable data support for subsequently strengthening directed networks.

[0043] In this embodiment, the property rights connection degree between each node in the undirected network is calculated, and the property rights connection degree is accumulated according to a preset weight onto the paths with corresponding node relationships in the directed network to generate a multi-source data reinforcement network, including: Traverse each directed path in the directed network and identify the starting and ending nodes corresponding to the directed path; Using the starting node and the ending node as search criteria, retrieve the corresponding connection records in the property rights connection matrix; If a corresponding contact record is found, the contact frequency in the contact record is extracted as the property rights connection strength. The property rights connection strength is multiplied by a preset strengthening weight coefficient to obtain the strengthening increment value. The reinforcement increment is accumulated onto the corresponding directed path, and the edge values ​​of the corresponding directed path are updated until all directed paths have been traversed, thus obtaining the multi-source data reinforcement network.

[0044] As an example, in the specific implementation process, the directed connection network is strengthened through undirected headquarters-branch and investment connections. Since the supply direction in headquarters-branch enterprises is undirected and the supply relationship is preset to a value of 0.5, it is only necessary for the headquarters enterprise, branch enterprise in the headquarters-branch connection and the purchasing enterprise and purchased enterprise in the purchasing connection to be of the same category (regardless of direction) to multiply the connection degree in the headquarters-branch connection by 0.5 and add it to the purchasing connection degree. Taking the headquarters-branch relationship as an example, the specific calculation formula is as follows: When any row i in the headquarters-branch category relationship matrix ZF_LB and any row j in the procurement category relationship matrix CG_LB satisfy the matching conditions: gbfl_zf_lb(i)=gbfl_cg_lb(j) and gbfl_bzf_lb(i)=gbfl_bcg_lb(j), or gbfl_zf_lb(i)=gbfl_bcg_lb(j) and gbfl_bzf_lb(i)=gbfl_cg_lb(j), then update the procurement relationship degree: lx_cg_lb(j)=lx_cg_lb(j)+0.5*lx_zf_lb(i).

[0045] Based on the aforementioned case data, the directed connection network CG_LB_Final is strengthened using the headquarters-branch connection ZF_LB: For [a, b, 2] in CG_LB, find the corresponding category relationship [a, b, 2] in ZF_LB, calculate the increment 0.5*2=1, and the relationship degree after the update is 2+1=3; For [d, e, 1] in CG_LB, find the corresponding category relationship [d, e, 1] in ZF_LB, calculate the increment 0.5*1=0.5, and the updated relationship degree is 1+0.5=1.5; For [d, c, 1] in CG_LB, find the corresponding category relationship [d, c, 1] in ZF_LB, calculate the increment 0.5*1=0.5, and the updated relationship degree is 1+0.5=1.5.

[0046] The final generated procurement category relationship matrix CG_LB is updated as follows: [a, b, 3, 1, 1] a, c, 5, 1, 1 d, e, 1.5, 1, 1 d, c, 1.5, 1, 1] like Figure 2 As shown, Figure 2 This is an updated procurement category connection network diagram provided in an embodiment of the present invention. In this case, the network diagram not only includes explicit procurement flows, but also incorporates implicit property rights association strength, making the connections between nodes more stable and realistic.

[0047] Understandably, in the industrial sector, headquarters-branch and investment connections only indicate ownership ties between industries and do not directly express whether there are clear upstream and downstream relationships. However, according to the logic of industrial cluster development, companies generally expand into related upstream and downstream industries during their expansion, resulting in ambiguous supply relationships. Therefore, this paper adopts a method that reinforces three types of industrial connection data: headquarters-branch, investment, and procurement. By using headquarters-branch and investment data to strengthen the edge values ​​of the directed network graph, it can quantify the strength of supply relationships in the industrial sector, compensate for the lack of sufficient industrial connection data sources, and obtain a highly interpretable industrial cluster entity connection network.

[0048] In this embodiment, all node paths in the multi-source data enhancement network are traversed, the cumulative connection strength of each node path is calculated, and the path with the highest cumulative connection strength is determined as the core industrial chain of the industrial cluster, including: Based on the multi-source data enhancement network, a node connectivity matrix is ​​constructed; the numerical values ​​of the elements in the node connectivity matrix represent the connection strength between the corresponding nodes. The values ​​of the elements in the node connectivity matrix are inverted to construct a path calculation matrix for path search. All nodes in the multi-source data augmentation network are traversed sequentially as intermediate nodes. For any pair of starting and ending nodes, the sum of the path value from the starting node to the intermediate node and the path value from the intermediate node to the ending node in the path calculation matrix is ​​calculated. Determine whether the sum of the path values ​​is less than the current path value from the starting node to the ending node recorded in the path calculation matrix. If so, update the current path value with the sum of the path values ​​until all nodes have been traversed. After traversing the path, obtain the absolute value of the path value between each pair of starting and ending nodes in the path calculation matrix. Determine the absolute value as the cumulative connection strength of the corresponding node path, and determine the node path with the largest cumulative connection strength as the core industrial chain of the industrial cluster.

[0049] As an example, the enterprise category nodes in the multi-source data augmentation network are first numbered sequentially as {1,2,3,...,i}, where the numbers represent enterprise category numbers. There are a total of i enterprise category nodes. A connection degree matrix D between the nodes is then constructed. D= [d(1,1),d(1,2),d(1,3)....d(1,j) d(1,1),d(1,2),d(1,3)....d(1,j) ... d(i,1),d(1,2),d(1,3)....d(i,j)], Here, d(i,j) represents the procurement connection degree between enterprise nodes i and j. If there is no connection between the nodes, it is assigned a value of +∞; if i=j, then d(i,j)=+∞. To use the Floyd algorithm to calculate the shortest path between each vertex in graph G=(V,E), two matrices are needed. The element a[i][j] in matrix S represents the distance from vertex i to vertex j; the element b[i][j] in matrix P represents the vertices represented by the values ​​recorded in b[i][j] that vertex i passes through on its way to vertex j. To facilitate the shortest path calculation in the algorithm, the values ​​in the connection degree matrix D need to be inverted, i.e., a[i][j]=-d(i,j) (if d(i,j)=+∞, then no inversion is needed, a[i][j]=d(i,j)). The shortest path calculated in this way is actually the path with the highest connection degree. Assuming there are N vertices in graph G, matrices D and P need to be updated N times. Initially, the distance of vertex a[i][j] in matrix D is the weight between vertex i and vertex j; the value of matrix P is the value of j of vertex b[i][j]. Next, matrix D is updated N times: during the k-th update, if the distance of a[i][j] is greater than a[i][k-1] + a[k-1][j] (i.e., the distance between i and j via the k-th vertex is shorter), then a[i][j] is updated to a[i][k-1] + a[k-1][j], and b[i][j] is also updated to b[i][k-1]. After N updates, the operation is complete.

[0050] Finally, based on the Floyd algorithm calculation results, the shortest distance matrix D between each node is obtained. The global minimum value min(D(i,j)) is searched in matrix D, and the path between the corresponding two nodes is searched out through matrix P to find L(i,j). L(i,j) is the path with the highest connectivity in the industrial cluster network, which is the core industrial chain.

[0051] Understandably, this embodiment flexibly utilizes the classic scenario of using the Floyd algorithm to find the minimum path in a directed network graph. By inverting the edge values ​​in the network graph (a[i][j]=-d(i,j)), the computational logic of the Floyd algorithm is successfully adapted, ensuring that the shortest path found globally in the calculation results corresponds precisely to the path with the highest industry connectivity. This method can automatically retrieve the combination of node pairs with the strongest connectivity globally, thereby objectively obtaining the core industrial chain in the directed network of the industrial cluster. It effectively avoids the subjectivity brought about by manually predefined industrial chain structures and achieves precise quantitative analysis driven by data.

[0052] To further verify the effectiveness of this method, we will take the automotive industry cluster as an example. Figure 3 As shown, Figure 3This is a schematic diagram of the core industry paths in the directed network of the automotive industry cluster provided in this embodiment of the invention. Using the method described in this embodiment, a directed connection network of the automotive industry cluster was constructed. In the network, new energy vehicle manufacturing, gasoline and diesel vehicle manufacturing, etc., serve as core nodes. After strengthening through multi-source data, multiple high-strength industry connection paths were identified. Through traversal calculation using the Floyd algorithm, a core industrial chain path was extracted as follows: [Pump and Vacuum Equipment Manufacturing] -- (Connection Strength 22) --> [Automotive Engine Manufacturing] -- (Connection Strength 26) --> [Gasoline and Diesel Vehicle Manufacturing] -- (Connection Strength 19) --> [New Car Retail]. This path clearly reflects the complete core chain from key components (pumps / engines) to vehicle manufacturing and downstream sales. Furthermore, the connection strengths of each link (such as 22, 26, and 19) are quantitative results that integrate procurement and property rights data, providing intuitive data support for the strengthening and supplementation analysis of the industrial cluster's supply chain.

[0053] Example 3 like Figure 4 As shown, this embodiment proposes an industrial cluster industrial chain construction device, which is applied to the industrial cluster industrial chain construction method described in the above embodiment, including: acquisition module 100, first construction module 200, second construction module 300, first calculation module 400 and second calculation module 500.

[0054] The acquisition module 100 is used to acquire supply-related data and property rights-related data of enterprises in the industry category to which the target industrial cluster belongs; the first construction module 200 is used to construct a directed connection network based on the supply-related data; the second construction module 300 is used to construct an undirected connection network based on the property rights-related data; the first calculation module 400 is used to calculate the property rights connection degree between each node in the undirected connection network, and accumulate the property rights connection degree according to a preset weight to the path with corresponding node relationship in the directed connection network to generate a multi-source data reinforcement network; the second calculation module 500 is used to traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

[0055] It should be noted that the foregoing explanation of the embodiment of the industrial cluster industrial chain construction method also applies to the industrial cluster industrial chain construction device of this embodiment, and will not be repeated here.

[0056] Example 4 Figure 5 This is a schematic diagram of the structure of the electronic device 600 provided in this embodiment. The electronic device 600 includes: a memory 601, a processor 602, and a computer program stored in the memory 601 and executable on the processor 602.

[0057] When the processor 602 executes the program, it implements the industrial cluster supply chain construction method provided in the above embodiments.

[0058] Furthermore, the electronic device 600 also includes a communication interface 603 for communication between the memory 601 and the processor 602.

[0059] The memory 601 may include high-speed RAM (Random Access Memory) and may also include non-volatile memory, such as at least one disk storage.

[0060] If the memory 601, processor 602, and communication interface 603 are implemented independently, then the communication interface 603, memory 601, and processor 602 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0061] Optionally, in a specific implementation, if the memory 601, processor 602, and communication interface 603 are integrated on a single chip, then the memory 601, processor 602, and communication interface 603 can communicate with each other through an internal interface.

[0062] Processor 602 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of the present invention.

[0063] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for constructing an industrial cluster supply chain.

[0064] In the description of this specification, the references to the terms "an embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0065] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0066] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0067] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (FPGAs), field-programmable gate arrays (FPGAs), etc.

[0068] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When executed, the program includes one or a combination of the steps of the method embodiments.

[0069] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for constructing an industrial cluster supply chain, characterized in that, include: Obtain supply-related data and ownership-related data of enterprises in the industry category to which the target industrial cluster belongs; A directed connection network is constructed based on the aforementioned supply-related data; Construct an undirected network based on the aforementioned property rights association data; Calculate the property rights connection degree between each node in the undirected network, and accumulate the property rights connection degree to the path with corresponding node relationship in the directed network according to the preset weight to generate a multi-source data reinforcement network; Traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

2. The method for constructing an industrial cluster supply chain according to claim 1, characterized in that, The acquisition of supply-related data and ownership-related data of enterprises in the target industrial cluster includes: Obtain descriptive text related to the target industry cluster, and extract keyword datasets from the descriptive text using a pre-defined large language model; The keyword dataset is matched with the preset national economic industry classification codes to determine the set of industry categories involved in the target industrial cluster; Obtain a list of surviving companies within the target region based on the aforementioned set of industry categories; The information on the related initiating companies, related receiving companies, and contact frequency of the companies in the list of existing companies is obtained as the property rights association data, and the information on the purchasing companies, the purchasing companies, and the purchasing frequency is obtained as the supply association data.

3. The method for constructing an industrial cluster supply chain according to claim 2, characterized in that, Constructing a directed connection network based on the aforementioned supply-related data includes: Quantitative processing is performed on supply-related data, including information on purchasing companies, information on companies being purchased, and purchase frequency, to construct a procurement relationship matrix; Based on the preset industry classification standards, the purchasing enterprise information and purchased enterprise information in the purchasing connection matrix are classified, and the purchasing frequency of records belonging to the same purchasing industry classification and the same purchased industry classification is accumulated to generate a directed connection network.

4. The method for constructing an industrial cluster supply chain according to claim 3, characterized in that, After generating the directed connection network, the method further includes: Traverse the directed relationship network. If a record has the same industry classification for purchasing as the industry classification for the purchased item, remove that record from the directed relationship network. Check if there are bidirectional connection records in the directed connection network that represent the purchasing industry category and the purchased industry category, respectively. If so, calculate the difference in purchasing frequency after summing the two directions, retain the record of the direction with the larger summed purchasing frequency, and update the purchasing connection degree of that record to the difference. If the summed purchasing frequencies of the two directions are the same, remove the bidirectional connection record to obtain the optimized directed connection network.

5. The method for constructing an industrial cluster supply chain according to claim 4, characterized in that, Constructing an undirected network based on the aforementioned property rights association data includes: The property rights association data, including information on related initiating companies, related receiving companies, and contact frequency, are quantified to construct a property rights association matrix. Based on the preset industry classification standards, the information of the associated initiating enterprises and the information of the associated receiving enterprises in the property rights connection matrix are mapped to the corresponding industry classifications, thereby generating an undirected connection network.

6. The method for constructing an industrial cluster supply chain according to claim 5, characterized in that, Calculate the property rights connection degree between each node in the undirected network, and accumulate the property rights connection degree according to a preset weight to the paths with corresponding node relationships in the directed network to generate a multi-source data reinforcement network, including: Traverse each directed path in the directed network and identify the starting and ending nodes corresponding to the directed path; Using the starting node and the ending node as search criteria, retrieve the corresponding connection records in the property rights connection matrix; If a corresponding contact record is found, the contact frequency in the contact record is extracted as the property rights connection strength. The property rights connection strength is multiplied by a preset strengthening weight coefficient to obtain the strengthening increment value. The reinforcement increment is accumulated onto the corresponding directed path, and the edge values ​​of the corresponding directed path are updated until all directed paths have been traversed, thus obtaining the multi-source data reinforcement network.

7. The method for constructing an industrial cluster supply chain according to claim 5, characterized in that, Traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and identify the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster, including: Based on the multi-source data enhancement network, a node connectivity matrix is ​​constructed; the numerical values ​​of the elements in the node connectivity matrix represent the connection strength between the corresponding nodes. The values ​​of the elements in the node connectivity matrix are inverted to construct a path calculation matrix for path search. All nodes in the multi-source data augmentation network are traversed sequentially as intermediate nodes. For any pair of starting and ending nodes, the sum of the path value from the starting node to the intermediate node and the path value from the intermediate node to the ending node in the path calculation matrix is ​​calculated. Determine whether the sum of the path values ​​is less than the current path value from the starting node to the ending node recorded in the path calculation matrix. If so, update the current path value with the sum of the path values ​​until all nodes have been traversed. After traversing the path, obtain the absolute value of the path value between each pair of starting and ending nodes in the path calculation matrix. Determine the absolute value as the cumulative connection strength of the corresponding node path, and determine the node path with the largest cumulative connection strength as the core industrial chain of the industrial cluster.

8. An industrial cluster supply chain construction device, characterized in that, include: The acquisition module is used to acquire supply-related data and property rights-related data of enterprises in the industry category to which the target industrial cluster belongs; The first construction module is used to construct a directed connection network based on the supply-related data; The second construction module is used to construct an undirected network based on the property rights association data; The first calculation module is used to calculate the property rights connection degree between each node in the undirected network, and to accumulate the property rights connection degree to the path with corresponding node relationship in the directed network according to the preset weight, so as to generate a multi-source data reinforcement network. The second calculation module is used to traverse all node paths in the multi-source data reinforcement network, calculate the cumulative connection strength of each node path, and determine the path with the highest cumulative connection strength as the core industrial chain of the industrial cluster.

9. An electronic device, characterized in that, The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the operations performed by the industrial cluster supply chain construction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that is executed by a processor as described in any one of claims 1 to 7, according to the method for constructing an industrial cluster supply chain.

Images