A regional industry node value evaluation and enterprise recommendation method

Abstract

The application discloses a regional industry node value evaluation and enterprise recommendation method, and the recommendation method comprises the following steps: S1, multi-source data acquisition and fusion; S2, using multi-source fusion data, constructing an industry dependence model, and calculating the quantitative dependence of regional industry nodes from four dimensions of capital relationship, technical cooperation, talent flow and transaction relationship; S3, constructing a strong chain industry node identification model; S4, constructing a chain extension and chain supplement industry node identification model; S5, after completing the industry node marking based on the strong chain, chain extension and chain supplement identification model, the possibility of target enterprise correlation reaching and the enterprise growth characteristics are deduced in combination with the industry dependence model, and a targeted enterprise recommendation method is constructed. The closed-loop decision mechanism from the industry node value evaluation to the enterprise recommendation is constructed, and the targeting of the enterprise recommendation and investment work is improved.

Description

Technical Field

[0001] This invention relates to the field of regional industrial node value assessment, and in particular to a method for assessing the value of regional industrial nodes and recommending investment enterprises. Background Technology

[0002] In other digital investment promotion platforms, expert experience is used to map the industrial chain, and publicly available data such as enterprise intellectual property rights and business scope are used to establish connections between enterprises and nodes in the industrial chain. Based on the industrial chain map of already linked enterprises, links with few or no related enterprises in the region are selected as recommended nodes for supply chain expansion. However, there is currently no research or practice on developing differentiated development strategies for strengthening, extending, and supplementing the industrial chain based on multi-dimensional data analysis of inter-regional interdependencies, combining factors such as capital penetration, technological collaboration, talent mobility, and supply chain topology.

[0003] In practical work, the following shortcomings exist in the application of existing technologies: (1) The granularity of macro data is too coarse, making it difficult to map to the micro-level main industrial links. Macro statistical data can usually only be accurate to the major categories of the national economy, making it difficult to drill down and correspond to specific micro-level nodes in the industrial chain; at the same time, statistical data has a lag of 2-3 years, which cannot reflect the real-time dynamics of the industry. (2) There is a lack of quantitative indicators for analyzing cross-regional industrial collaboration. Existing systems cannot accurately measure the true dependence of a local industrial link on a specific surrounding region. Due to the lack of real transaction and collaboration data support between micro-enterprises, local industrial planning is prone to homogeneous competition, making it difficult to achieve scientific and differentiated development. (3) The investment promotion strategy is not targeted enough. When recommending enterprises, existing investment promotion systems mostly use lists, tags or industries for filtering, failing to accurately match with the development of the local supply chain, resulting in insufficient targeting of investment promotion work. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method for regional industrial node value assessment and investment enterprise recommendation that overcomes or at least partially solves the above problems.

[0005] According to one aspect of the present invention, a method for assessing the value of regional industrial nodes and recommending enterprises for investment is provided, the recommendation method comprising: Step S1: Acquisition and fusion of multi-source data; Step S2: Utilize multi-source fusion data to construct an industry interdependence model, and calculate the quantitative interdependence of industrial nodes between regions from four dimensions: capital relations, technological collaboration, talent flow, and transaction relations; Step S3: Construct a strong industrial chain node identification model; Step S4: Construct an identification model for industry nodes in the extended and supplementary supply chains; Step S5: After completing the industry node labeling based on the strong chain, extended chain, and supplementary chain identification model, combine the industry dependence model to infer the likelihood of reaching the target enterprise and the enterprise's growth characteristics, and construct a targeted enterprise recommendation method for investment attraction.

[0006] Optionally, step S1: multi-source data acquisition and fusion specifically includes: Data Acquisition: Acquire all required data in a compliant manner, including external data such as industry chain knowledge graphs, enterprise business registration data, enterprise annual report data, enterprise announcement data, publicly available judicial data, senior management appointment data, listing data, enterprise recruitment data, bidding and winning data, private equity and venture capital data, honor and qualification data, intellectual property data, and macroeconomic indicator data. Among them, the industry chain knowledge graph data includes industry chain graphs, upstream and downstream relationships, and enterprises associated with each industry node. Data preprocessing: Investigate the data situation, clean up missing, duplicate, and invalid data, design standard table structures and data formats, and integrate, manage, and standardize data from multiple sources; Data fusion and entity alignment: Design unique identifiers for the corporate entities and natural persons involved in the data, and associate and align data involving the same entity from multiple sources with unique identifiers.

[0007] Optionally, the capital relationship indicator is used to quantify the degree of capital penetration of the target region into specific industrial nodes in the local region; By integrating multi-source data such as business registration, annual business reports, financial reports of listed companies, and investment and financing events published by private equity and venture capital firms, a quantitative calculation model of capital relationships with controlling stake premium is constructed.

[0008] Optionally, the technology collaboration index is used to quantify the degree of dependence of specific industrial nodes in the local area on the target area in terms of technology research and development, intellectual property sharing and technology transfer; the data sources include publicly available patent / software copyright joint application data, patent transfer announcement data, and publicly available standard drafting unit data.

[0009] Optionally, the talent mobility index is used to quantify the dependence of specific industry nodes in a local area on the target region in terms of the historical tenure, cross-appointment, and high-end R&D talent of corporate executives; data sources include key personnel data from business registration, the resumes of directors, supervisors, and senior executives disclosed in the financial reports of listed companies, and publicly available patent inventor data.

[0010] Optionally, the transaction relationship indicator is used to quantify the degree of dependence of a specific industry node in the local area on the target area in transaction behavior; the data sources include bidding data, prospectuses and corporate financial reports, announcements of major corporate contracts, and publicly available judicial judgments.

[0011] Optionally, step S3: constructing a strong industrial chain node identification model specifically includes: By integrating the logic of macroeconomic operation with big data on enterprise micro-operations, strong chain nodes can be automatically identified by calculating the characteristic scores of various industrial nodes. For specific areas industrial chain nodes Its strong chain value comprehensive evaluation index consists of scores from four dimensions: node scale and quality, operational activity, innovation index, and micro-location quotient. set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0012] Optionally, step S4: constructing the industry node identification model for extending and supplementing the industrial chain specifically includes: The chain extension and supplementary chain node identification model uses the industrial chain knowledge graph and the aforementioned industrial dependence model and strong chain node identification model as inputs to identify industrial nodes with high synergy potential and located in areas with little regional competition.

[0013] Optionally, step S5: After completing the industry node labeling based on the strong chain, extended chain, and complementary chain identification model, and combining the industry dependency model to infer the likelihood of reaching the target enterprise and the enterprise's growth characteristics, a targeted investment attraction enterprise recommendation method is constructed, specifically including: Target investment promotion nodes The steps to recommend target companies: Initial screening of candidate companies: In the industry chain knowledge graph, all companies mapped to nodes are screened. Excluding companies already registered in this region based on their business registration address. For companies that register or establish significant branches, an initial set of candidate companies is generated. ; Potential for Relationship Outreach Assessment: Based on the industry dependence of the target company's location on the local industry, identify key investment attraction areas to increase the success rate of reaching out through existing local businesses in these areas; define candidate companies. The registered location is outside the province or city. ,area At the node Up to local area The overall dependency ratio is ; Define Enterprise Relationship reach probability score Enterprise quality assessment: Based on publicly available honor and qualification lists and judicial negative list data, candidate enterprises are given a simple quality score; Define Enterprise Quality score in As the basic score for enterprises, For positive qualifications, if a company is on the positive list, it will be given the corresponding bonus points according to its qualification level. The negative list will deduct points; if a company hits a major risk data point, its score will be lowered. Overall Recommendation Ranking: After normalizing the above two scores, the final investment recommendation score is calculated. in , These are preset weighting coefficients, which typically prioritize ensuring enterprise quality. ; according to Sort the candidate companies in descending order and truncate the top few. Each company generates a targeted list of potential investment partners; For high-scoring companies on the list, extract publicly available information about the target companies and associate the contact paths.

[0014] This invention provides a method for regional industrial node value assessment and investment enterprise recommendation. The recommendation method includes: Step S1: multi-source data acquisition and fusion; Step S2: using multi-source fused data to construct an industrial interdependence model, calculating the quantitative interdependence of industrial nodes between regions from four dimensions: capital relations, technological collaboration, talent mobility, and transaction relations; Step S3: constructing a strong-chain industrial node identification model; Step S4: constructing an extended-chain and complementary-chain industrial node identification model; Step S5: after completing the industrial node labeling based on the strong-chain, extended-chain, and complementary-chain identification models, combining the industrial interdependence model to infer the likelihood of reaching target enterprises and the growth characteristics of enterprises, and constructing a targeted investment enterprise recommendation method. This establishes a closed-loop decision-making mechanism from industrial node value assessment to investment enterprise recommendation, improving the targeting of investment attraction efforts.

[0015] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for assessing the value of regional industrial nodes and recommending investment enterprises, provided as an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0019] The terms "comprising" and "having," and any variations thereof, in the specification, embodiments, claims, and drawings of this invention are intended to cover non-exclusive inclusion, such as including a series of steps or units.

[0020] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0021] like Figure 1 As shown, a method for regional industrial node value assessment and enterprise recommendation for investment attraction is provided, the recommendation method including: Step S1: Acquisition and fusion of multi-source data; Step S2: Utilize multi-source fusion data to construct an industry interdependence model, and calculate the quantitative interdependence of industrial nodes between regions from four dimensions: capital relations, technological collaboration, talent flow, and transaction relations; Step S3: Construct a strong industrial chain node identification model; Step S4: Construct an identification model for industry nodes in the extended and supplementary supply chains; Step S5: After completing the industry node labeling based on the strong chain, extended chain, and supplementary chain identification model, combine the industry dependence model to infer the likelihood of reaching the target enterprise and the enterprise's growth characteristics, and construct a targeted enterprise recommendation method for investment attraction.

[0022] I. Acquisition and Fusion of Multi-Source Data 1. Data Acquisition. Acquire all required data through compliant means, including external data such as industry chain knowledge graphs, enterprise registration data, enterprise annual report data, enterprise announcement data, publicly available judicial data, senior management appointment data, listing data, enterprise recruitment data, bidding and winning data, private equity and venture capital data, honor and qualification data, intellectual property data, and macroeconomic indicator data. Among them, the industry chain knowledge graph data includes industry chain graphs, upstream and downstream relationships, and enterprises associated with each industry node.

[0023] 2. Data Preprocessing. Investigate the data situation, clean up missing, duplicate, and invalid data, design standard table structures and data formats, and integrate, manage, and standardize data from multiple sources.

[0024] 3. Data fusion and entity alignment. For the business entities and natural persons involved in the data, unique identifiers are designed to associate and align data from multiple sources involving the same entity, ensuring that relevant data can be correctly linked through unique identifiers under statistical standards.

[0025] II. Construction of Industry Dependence Model By utilizing multi-source fusion data, an industry interdependence model is constructed to calculate the quantitative interdependence of industrial nodes between regions from four dimensions: capital relations, technological collaboration, talent mobility, and transaction relations.

[0026] 1. Capital Relationship Indicators Capital relationship indicators are used to quantify the degree of capital penetration of a target region into specific industry nodes within the local area. By integrating multi-source data such as business registration, annual business reports, financial reports of listed companies, and investment and financing events disclosed by private equity and venture capital firms, a quantitative calculation model of capital relationships with a controlling stake premium is constructed.

[0027] (1) Parameter definition Let the local area be The target comparison area is Specific nodes in the industrial chain are .

[0028] set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0029] set up For the place of registration in the region The collection of all capital-exporting entities (including enterprises, investment institutions, etc.).

[0030] set up It refers to the collection of entities that export all capital (i.e., investors nationwide).

[0031] (2) Effective capital injection function definition For entities For enterprises The effective capital injection valuation. Because data from multiple sources may have statistical overlap (e.g., the financing amount of a private equity / venture capital investment event may also be reflected as an increase in registered capital in business registration), a priority coverage deduplication alignment function is used. Perform the calculation: in The equity value is calculated based on business registration and annual business reports (the amount of shares held converted from subscribed / paid-in capital or shareholding ratio). This refers to the historical cumulative investment and financing amount extracted from the private equity and venture capital database. The market value of the shares held is calculated based on the data of the top ten shareholders / circulating shares disclosed in the financial reports of listed companies, and converted according to the calculation benchmark date.

[0032] Deduplication and alignment functions The logic is that when the same investment entity When records exist in multiple data sources, the market capitalization based on the listed company's financial statements is used. As the top priority, private equity venture capital amount The second priority is the converted value of business registration. The value is assigned to the third priority level.

[0033] (3) Controlling premium coefficient To measure the control capital has over industry nodes, a controlling premium coefficient is introduced. Let... For entities For enterprises The actual shareholding ratio.

[0034] when Or, if the actual controller is identified through penetration, it is determined to be an absolute controlling shareholder, and a system is set up. (e.g., taking the value 1.5), when At that time, it was determined to be a shareholder with significant influence, and a system was established. (e.g., taking the value 1.2), when At that time, it was determined to be a general financial investment and set up .

[0035] (4) Calculation of capital dependence index area node For the region capital dependence The calculation model is as follows: Where the molecule is a region Capital for the region node The total effective capital injection of all enterprises within the region (including controlling premium), with the denominator being the region. node The total amount of effective capital injections (including controlling premium) absorbed by all enterprises nationwide from all regions. (Indicator value) The larger the value, the more the local industry node is dependent on the target region in terms of capital relations.

[0036] 2. Technological Collaboration Indicators The technology collaboration index is used to quantify the degree to which specific industry nodes in a local area rely on a target region in terms of technology research and development, intellectual property sharing, and technology transfer. Data sources include publicly available patent / software copyright joint application data, patent transfer announcement data, and publicly available standard drafting unit data.

[0037] (1) Parameter definition Let the local area be The target comparison area is Specific nodes in the industrial chain are .

[0038] set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0039] set up For the place of registration in the region The collection of all innovation entities (including enterprises, universities, research institutes, etc.).

[0040] set up It refers to the entire collection of innovation entities (i.e., innovation entities nationwide).

[0041] (2) Jointly develop functions definition For entities With entity The overall valuation of joint research and development, including jointly applied intellectual property rights and jointly drafted standards: in For entities and The total number of jointly authored technological achievements. For the first The weight of each technological achievement is determined by its type: national standards / international standards have the highest weight (e.g., 2.0), industry standards / invention patents have the second highest weight (e.g., 1.0), group standards / utility models have the lowest weight (e.g., 0.5), and design patents and software copyrights have the lowest weight (e.g., 0.2). For the first The time-related decline factor of the achievement is determined using an exponential decline model. ,in The time difference (in years) between the application date or publication date of the achievement and the current evaluation benchmark date. It is a decreasing constant (e.g., a value of 0.1 indicates that the longer the technology has been published, the lower its contribution to the current technology synergy index).

[0042] (3) Technology transfer function definition For entities With entity The value of technological achievements transferred and commercialized between them. Calculated based on publicly available patent change and licensing records. in This refers to the total number of patent transfer or licensing events between the two parties. These are the type weight and the time-decrease factor of the transferred patent, respectively. The transfer depth coefficient is as follows: the highest value is for complete transfer of patent rights (e.g., 1.0), followed by exclusive license (e.g., 0.8), then sole license (e.g., 0.5), and the lowest value is for ordinary license (e.g., 0.3).

[0043] (4) Calculation of the technology collaboration interdependence index Define entity and Comprehensive technical synergy score in , The pre-defined weighting coefficients are used to measure the contribution of joint R&D and technology sales to the collaborative score.

[0044] area node For the region Technical Co-dependence Indicators The calculation model is Where the molecule is a region node Domestic enterprises and regions The sum of the comprehensive technological collaboration scores of the innovation entities. The denominator is the region. node The sum of the comprehensive technological collaboration scores between domestic enterprises and all innovation entities nationwide. (Indicator value) The larger the value, the more the local industry node relies on the target region for technology research and development and technology input.

[0045] 3. Talent Mobility Indicators The talent mobility indicator is used to quantify the dependence of specific industry nodes within a local area on a target region in terms of the historical tenure, cross-appointment, and high-end R&D talent of corporate executives. Data sources include key personnel data from business registration, the resumes of directors, supervisors, and senior executives disclosed in listed companies' financial reports, and publicly available patent inventor data.

[0046] (1) Parameter definition Let the local area be Target comparison area Specific nodes in the industrial chain are .

[0047] set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0048] set up For the place of registration in the region The collection of all enterprises and research institutions.

[0049] set up It is a collection of entities nationwide.

[0050] set up For corporate entities Registered capital (or paid-in capital / latest valuation, taking the optimal value based on the alignment results of multi-source data, unit: RMB 10,000).

[0051] (2) Historical talent mobility function definition For entities (Registered in the region) From entity (Registered in the region) This involves acquiring historical talent transfer value. NLP technology is used to identify the past employment information of senior executives within companies, as well as cross-company situations involving inventors with the same name and in the same technical field.

[0052] in: For those who were in the physical Currently employed and joining the entity A group of people . For personnel The job title and rank weighting factor is set as follows: Founder / CEO / CTO has the highest weight (e.g., 1.0), followed by core technical personnel / first inventor of patents (e.g., 0.8), and general directors, supervisors and senior executives (e.g., 0.5). As a time-decrease factor: ,in For personnel from The time difference between leaving the job and the current assessment benchmark date (in years). It is a decreasing constant.

[0053] (3) Cross-appointment collaboration function definition For entities With entity Synergistic value arising from overlapping personnel roles: in: This refers to the group of natural persons who currently hold director, supervisor, senior management, or core technical positions in two entities simultaneously. . The highest rank weight for personnel holding concurrent positions.

[0054] (4) Firm size factor Considering that large enterprises have a stronger technology spillover effect and industrial influence, the impact of their talent mobility on the dependence of industrial nodes is measured using the following model. This model uses the natural logarithm to smooth the registered capital. The smoothing constant (e.g., the natural constant) This is to avoid abnormal extreme values ​​in indicators caused by excessively large or small registered capital of a single enterprise.

[0055] (5) Calculation of talent mobility dependence index Define entity For entities The overall talent dependence score is as follows, among which , These are the normalized weighting coefficients.

[0056] area node For the region Dependence on talent mobility The calculation model is Where the molecule is a region node Domestic enterprises to the region The overall talent pool of an entity depends on the sum of its scores, with the denominator being the region. node The sum of the comprehensive talent dependence scores of domestic enterprises on all entities nationwide. This indicator value... The higher the value, the more the local industry node relies on the target region for talent supply.

[0057] 4. Transaction Relationship Indicators The transaction relationship indicator is used to quantify the degree to which a specific industry node in a local area depends on a target area in its transaction activities. Data sources include bidding data, prospectuses and corporate financial reports, announcements of major corporate contracts, and publicly available judicial documents.

[0058] (1) Parameter definition Let the local area be The target comparison area is Specific nodes in the industrial chain are .

[0059] set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0060] set up For the place of registration in the region The collection of all enterprise entities.

[0061] set up It is a collection of business entities nationwide.

[0062] set up For entities With entity A collection of historical transaction events between them.

[0063] (2) Transaction event value function For any single transaction event Its basic transaction amount is Because some data sources (such as general corporate announcements and some judicial documents) only disclose the transaction relationship but not the specific amount, dynamic interpolation is used to fill in the gaps based on company size and industry: in The amount of the winning bid, the amount disclosed in the prospectus, or the amount involved in a case as stated in judicial documents. Registered capital of the enterprise (unit: RMB 10,000). This is the smoothing constant. This is the industry benchmark transaction coefficient (derived by fitting the average size of publicly available transaction data for this industry node). (3) Data source confidence and timeliness weight Different data sources reflect varying degrees of reliability in transaction relationships, and their influence should diminish over time. The effective collaborative value of a single transaction event is defined as... in Assign confidence weights to data sources. The rules are as follows: ongoing supplier relationships disclosed in prospectuses / audited financial statements have the highest weight (e.g., 1.2), followed by bidding and tendering data (e.g., 1.0), major contract announcements (e.g., 0.8), and sales contracts extracted from judicial documents have a lower weight (due to the risk of cooperation breakdown, the weight is reduced, e.g., 0.5). For the diminishing returns factor of transactions: ,in The time difference (in years) between the date of the transaction or the date of the document's publication and the present. It is a decreasing constant (e.g., a value of 0.2). (4) Calculation of the interdependence index of transaction relationship Define entity With entity Comprehensive transaction relationship score between area node For the region Dependence on transaction relationships The calculation model is as follows: Where the molecule is a region node Domestic enterprises and regions The sum of all transactions undertaken by the enterprise, with the denominator being the region. node The total number of transactions between domestic enterprises and all enterprises nationwide. (Indicator value) The higher the value, the more the local industry node relies on the target region for order acquisition or raw material / service procurement.

[0064] III. Construction of a Strong Chain Industry Node Identification Model This model attempts to avoid the limitations of overly coarse macroeconomic statistical data in traditional practices. It integrates the logic of macroeconomic operation with big data on enterprise micro-operations and automatically identifies strong chain nodes with comparative advantages and the potential to be further expanded and strengthened by calculating the characteristic scores of various industry nodes.

[0065] For specific areas industrial chain nodes Its strong chain value comprehensive evaluation index consists of scores from four dimensions: node scale and quality, operational activity, innovation index, and micro-location quotient. Let... For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

[0066] 1. Industry node scale and quality score This is used to quantify the absolute industrial size and high-quality enterprise concentration of the industry node in the local area.

[0067] (1) Enterprise quality weight Define enterprises by using data from honor and qualification ranking lists and publicly available judicial data. Quality weight If a company is included in a positive list (such as a specialized and innovative enterprise, a high-tech enterprise, or a single champion), then... (e.g., a value of 1.5); if the enterprise is on the negative list in publicly available judicial data (e.g., a dishonest judgment debtor, a party involved in a major tax violation case), set... (i.e., remove its size contribution); if none of them are hit, set .

[0068] (2) Size score index Define Enterprise The registered capital is The weighted score can be obtained by weighting the registered capital. 2. Score of operational activity of industry nodes Based on private equity and venture capital data and publicly tendered data, the industry nodes are calculated. The overall fundraising ability and business expansion activity of domestic enterprises in This is a collection of private equity, venture capital, and private placement financing events that have occurred in companies within this node over the past three years. For the amount of a single round of financing, This is a collection of bidding events in which companies within this node have participated as successful bidders over the past three years. This refers to the amount of a single winning bid. As a time-decreasing factor, financing or bidding events that are further removed from the current valuation date have a lower weighting in terms of their contribution to current business activity. , This is the preset smoothing parameter.

[0069] 3. Innovation Score of Industry Nodes The innovation score of an industry node is calculated by combining the intellectual property scale, growth rate, concentration, and R&D intentions of enterprises within the industry node. in This refers to the total number of valid patents and software copyrights owned by companies within this node. This represents the compound annual growth rate of intellectual property applications at this node over the past three years. This is the Herfindahl index for the distribution of intellectual property rights within this node. The larger the value, the more likely the innovation is to be contributed by multiple companies, indicating a vibrant innovation ecosystem within the industrial cluster. To act as a proxy variable for research intentions, we consider the proportion of job postings with research-related categories in publicly available recruitment data to the total number of job postings at that node.

[0070] 4. Micro-location quotient index Referring to the calculation method of macroeconomic location quotient, the degree of industrial agglomeration is calculated using publicly available micro-enterprise operating data.

[0071] Where the molecule is a region internal nodes The proportion of the total registered capital to the total registered capital of all industries in this region, with the denominator being nodes nationwide. The proportion of its total registered capital to the total registered capital of all industries nationwide.

[0072] 5. Strong Link Node Identification Model (1) Normalization function To eliminate the differences in the dimensions and magnitudes of the indicators across the four dimensions mentioned above, for any evaluation dimension... Set up local area The original score set of all industry nodes in this dimension is: The model uses a range normalization algorithm to map the raw scores of each node to... The dimensionless interval is used to obtain the normalized score. in and These are the minimum and maximum values ​​of the original scores for this evaluation dimension across all local nodes, respectively. This is a preset, minimal constant used to avoid computational errors.

[0073] (2) Strong chain comprehensive score Based on the normalized scores of each dimension, the node values ​​are calculated using a linear weighted model. Strong chain composite score in, , , , The preset decision weight coefficients, and satisfy the following conditions: The weighting coefficients can be dynamically adjusted based on regional industrial policy guidance and analytical practices. For example, when a specific node... When both of the following conditions are met ① Its original location quotient (Ensure it possesses a fundamental comparative advantage) ② Its overall score Located at the top of the local node scores in descending order (e.g., the first 15%) The industry nodes that meet the criteria are labeled with a strong link in the industrial chain, and this information is then input into the investment promotion recommendation model to automatically filter and push a list of high-value enterprises that match the attributes of that node nationwide.

[0074] IV. Construction of an Identification Model for Industrial Nodes in the Extension and Supplementation of the Industrial Chain The chain extension and supplementary chain node identification model uses the industrial chain knowledge graph and the aforementioned industrial dependence model and strong chain node identification model as inputs to identify industrial nodes with high synergy potential and located in areas with little regional competition.

[0075] 1. Parameter Definition Let the local area be The surrounding areas of the same level are collectively known as .

[0076] set up This is a directed graph of the industry chain topology, in which... For all industry nodes, This is the set of upstream and downstream edges connecting nodes.

[0077] Suppose the local region is calculated based on the "Strong Chain Industry Node Identification Model" in Section 3. The set of strong chain nodes is .

[0078] Set up surrounding area The set of strong chain nodes is .

[0079] set up For the region At the industry node Above and in the region The overall dependence is calculated based on the "Industry Dependence Model" in Section 2.

[0080] 2. Chain Node Identification Model The model is designed to extend upstream and downstream around existing strong local nodes, but avoids advantageous links in surrounding areas, in order to achieve differentiated extension of local industries.

[0081] In the map Traverse all nodes If and only if A node will be included in the candidate set of extension nodes if it meets all four of the following constraints: Condition 1 (Local non-strong chain constraint): This means that the node is not currently a strong node in the region and has room for further development.

[0082] Condition 2 (surrounding misalignment and exclusive constraint): This means that the node is not a strong chain node in any of its surrounding peer regions, thus avoiding homogeneous competition.

[0083] Condition 3 (Upstream and downstream constraints of the graph): , making or , i.e., target node It must be linked to a local strong link node. There is a direct upstream and downstream topological relationship.

[0084] Condition 4 (High Dependency Constraint): Calculate Local Strong Link Nodes Rank of dependence on all surrounding areas of the same level . If this area (or in front) This indicates that the strong chain node has significant room for expansion in the region.

[0085] For the extended chain nodes that meet the above conditions Calculate its extended chain recommendation score : The industry average gross profit margin for the target node is used to assess the potential revenue of the target node. , , These are the balancing weighting coefficients.

[0086] 3. Supply Chain Node Identification Model The model is constructed in a similar way to the extended chain, starting from the strong chain nodes in the surrounding areas and providing supporting facilities for the strong chain nodes in the surrounding areas to achieve cross-regional collaborative chain supplementation.

[0087] In the map Traverse all nodes If and only if A chain is included in the candidate set of complement chains only if it meets all four of the following constraints: Condition 5 (Local non-strong chain constraint): This means that the node is not currently a strong chain node in the region.

[0088] Condition 6 (surrounding misalignment and exclusive constraint): This means that the node is not a strong chain node in the surrounding area, thus avoiding homogeneous competition with the dominant industries in the surrounding area.

[0089] Condition 7 (Upstream and Downstream Constraints of the Graph): and , making or , i.e., target node It must be a strong link node in a certain surrounding area. The direct upstream and downstream links.

[0090] Condition 8 (Cross-region high dependency constraint): Calculate the strong chain nodes in the surrounding region. For the local area The overall dependence ranking, if the region This indicates that strong links in the surrounding industrial chain have a strong actual demand for local factor supply or supporting services.

[0091] For the chain completion node that meets the above conditions The system calculates its complementary chain recommendation score. : in The stronger the surrounding blockchain nodes, the higher the commercial value of local blockchain support. The industry average gross profit margin for the target node is used to assess the potential revenue of the target node. , , To balance the weight system.

[0092] 4. Automated tagging and differentiated investment attraction strategy output: The industry nodes that meet the criteria are labeled with a strong link in the industrial chain, and this information is then input into the investment promotion recommendation model to automatically filter and push a list of high-value enterprises that match the attributes of that node nationwide.

[0093] Targeted investment attraction enterprise recommendations After labeling industry nodes based on the strong chain, extended chain, and supplementary chain identification model, the second step, the "industry dependence model," is used to infer the likelihood of reaching target companies and their growth characteristics, thereby constructing a targeted investment attraction company recommendation method.

[0094] Target investment promotion nodes (Including strong chain, extended chain, and supplementary chain nodes), the steps for recommending target companies are as follows: 1. Initial screening of candidate companies In the industry chain knowledge graph, filter all nodes mapped to it. Excluding companies already registered in this region based on their business registration address. For companies that register or establish significant branches, an initial set of candidate companies is generated. .

[0095] 2. Assessment of Potential Outreach Based on the industry dependence of the target company's region on the local industry, identify key investment attraction areas to increase the success rate of reaching out to local businesses within those areas. Candidate companies are defined. The registered location is outside the province or city. ,area At the node Up to local area The overall dependency ratio is .

[0096] Define Enterprise Relationship reach probability score 3. Enterprise Quality Assessment Based on publicly available honor and qualification rankings and judicial negative list data, candidate companies are given a simple quality score.

[0097] Define Enterprise Quality score in The basic score for the enterprise (such as a logarithmic function or other scoring model based on registered capital or years of establishment). For positive qualifications, if a company is listed in positive categories such as "Specialized, Refined and Innovative Little Giant", "Unicorn", "Gazelle Enterprise", or "High-tech Enterprise", it will be given corresponding bonus points according to its qualification level. The negative list will deduct points if a company is listed as a "dishonest judgment debtor" or "serious illegal and dishonest enterprise."

[0098] 4. Overall Recommendation Ranking After normalizing the above two scores, the final investment recommendation score is calculated. in , These are preset weighting coefficients, which typically prioritize ensuring enterprise quality. .according to Sort the candidate companies in descending order and truncate the top few. The system generates a targeted list of investment attraction companies. For high-scoring companies on the list, publicly available information on these companies can be extracted using business registration data, successful bid data, and private equity / venture capital data. Alternatively, the system can explore related reach paths by analyzing the investment and transaction relationships of local companies.

[0099] Beneficial effects: 1. A multi-dimensional quantitative calculation method for regional industrial dependence is proposed to provide quantifiable data support for differentiated development and ecological complementarity among regions.

[0100] Existing technologies lack objective and quantitative indicators for measuring cross-regional industrial synergy, easily leading to homogeneous competition in local industrial planning. This method innovatively constructs a dependency calculation model from four dimensions: capital relations, technological synergy, talent mobility, and transaction relations. In the process of identifying nodes for extending and supplementing the industrial chain, the macro-concepts of differentiated development and ecological complementarity are transformed into quantifiable constraints. This automatically identifies and avoids advantageous links in surrounding areas, proactively seeking out uncontested areas with high synergy potential, thus achieving cross-regional industrial ecological complementarity.

[0101] 2. Apply the logic of macroeconomic operation to micro-enterprise big data calculation to overcome the problems of coarse granularity and data lag in traditional industry analysis.

[0102] Current industry analysis practices mainly rely on macro-level statistical yearbooks, which cannot be mapped to micro-level industry segments. This method integrates heterogeneous data from multiple sources, such as business registration, intellectual property, bidding, and private equity venture capital, to construct a quantitative assessment model for the value of industry nodes, including micro-level location savvy and business activity. This model drills down to specific micro-level industry chain nodes and introduces a time-decrease factor into the calculation to improve the accuracy and timeliness of industry node value assessment.

[0103] 3. Establish a closed-loop decision-making mechanism from industrial node value assessment to investment enterprise recommendation to improve the pertinence of investment promotion work.

[0104] Existing investment promotion systems often rely on simple keyword or company tag filtering, resulting in recommendations that lack deep integration with the local supply chain. This method, after automatically tagging nodes in the strong, extended, and supplementary supply chains, comprehensively ranks companies based on their quality and reachability. This not only ensures that recommended companies align with local industrial planning needs but also leverages knowledge graph mining to provide investment promotion personnel with reach paths based on historical transactions, capital, or talent connections, effectively addressing the problems of traditional investment promotion efforts being largely aimless and lacking focus.

[0105] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for assessing the value of regional industrial nodes and recommending enterprises for investment, characterized in that, The recommendation method includes: Step S1: Acquisition and fusion of multi-source data; Step S2: Utilize multi-source fusion data to construct an industry interdependence model, and calculate the quantitative interdependence of industrial nodes between regions from four dimensions: capital relations, technological collaboration, talent flow, and transaction relations; Step S3: Construct a strong industrial chain node identification model; Step S4: Construct an identification model for industry nodes in the extended and supplementary supply chains; Step S5: After completing the industry node labeling based on the strong chain, extended chain, and supplementary chain identification model, combine the industry dependence model to infer the likelihood of reaching the target enterprise and the enterprise's growth characteristics, and construct a targeted enterprise recommendation method for investment attraction.

2. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, Step S1: Multi-source data acquisition and fusion specifically includes: Data Acquisition: Acquire all required data in a compliant manner, including external data such as industry chain knowledge graphs, enterprise business registration data, enterprise annual report data, enterprise announcement data, publicly available judicial data, senior management appointment data, listing data, enterprise recruitment data, bidding and winning data, private equity and venture capital data, honor and qualification data, intellectual property data, and macroeconomic indicator data. Among them, the industry chain knowledge graph data includes industry chain graphs, upstream and downstream relationships, and enterprises associated with each industry node. Data preprocessing: Investigate the data situation, clean up missing, duplicate, and invalid data, design standard table structures and data formats, and integrate, manage, and standardize data from multiple sources; Data fusion and entity alignment: Design unique identifiers for the corporate entities and natural persons involved in the data, and associate and align data involving the same entity from multiple sources with unique identifiers.

3. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, The capital relationship indicator is used to quantify the degree of capital penetration of the target region into specific industrial nodes in the local area; By integrating multi-source data such as business registration, annual business reports, financial reports of listed companies, and investment and financing events published by private equity and venture capital firms, a quantitative calculation model of capital relationships with controlling stake premium is constructed.

4. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, The aforementioned technology collaboration indicators are used to quantify the degree of dependence of specific industrial nodes in the local area on the target region in terms of technology research and development, intellectual property sharing, and technology transfer. Data sources include publicly available patent / software copyright joint application data, patent transfer announcement data, and publicly available standard drafting unit data.

5. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, The talent mobility index is used to quantify the dependence of specific industry nodes in a local area on the target region in terms of the historical tenure, cross-appointment, and high-end R&D talent of corporate executives; the data sources include key personnel data from business registration, the resumes of directors, supervisors, and senior executives disclosed in the financial reports of listed companies, and publicly available patent inventor data.

6. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, The transaction relationship indicator is used to quantify the degree of dependence of a specific industry node in the local area on the target area in transaction behavior; the data sources include bidding data, prospectuses and corporate financial reports, announcements of major corporate contracts, and publicly available judicial judgments.

7. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, Step S3: Constructing a strong industry chain node identification model specifically includes: By integrating the logic of macroeconomic operation with big data on enterprise micro-operations, strong chain nodes can be automatically identified by calculating the characteristic scores of various industrial nodes. For specific areas industrial chain nodes Its strong chain value comprehensive evaluation index consists of scores from four dimensions: node scale and quality, operational activity, innovation index, and micro-location quotient. set up For the place of registration in the region And it is mapped to industry nodes A collection of corporate entities.

8. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, Step S4: Constructing the identification model for industry nodes in the extended and supplementary supply chains specifically includes: The chain extension and supplementary chain node identification model uses the industrial chain knowledge graph and the aforementioned industrial dependence model and strong chain node identification model as inputs to identify industrial nodes with high synergy potential and located in areas with little regional competition.

9. The method for regional industrial node value assessment and investment enterprise recommendation according to claim 1, characterized in that, Step S5: After labeling industry nodes based on the strong chain, extended chain, and supplementary chain identification models, the method of constructing a targeted investment attraction enterprise recommendation method is inferred by combining the industry dependency model to determine the likelihood of reaching target enterprises and their growth characteristics. Specifically, this includes: Target investment promotion nodes The steps to recommend target companies: Initial screening of candidate companies: In the industry chain knowledge graph, all companies mapped to nodes are screened. Excluding companies already registered in this region based on their business registration address. For companies that register or establish significant branches, an initial set of candidate companies is generated. ; Potential for Relationship Outreach Assessment: Based on the industry dependence of the target company's location on the local industry, identify key investment attraction areas to increase the success rate of reaching out through existing local businesses in these areas; define candidate companies. The registered location is outside the province or city. ,area At the node Up to local area The overall dependency ratio is ; Define Enterprise Relationship reach probability score Enterprise quality assessment: Based on publicly available honor and qualification lists and judicial negative list data, candidate enterprises are given a simple quality score; Define Enterprise Quality score in As the basic score for enterprises, For positive qualifications, if a company is on the positive list, it will be given the corresponding bonus points according to its qualification level. The negative list will deduct points; if a company hits a major risk data point, its score will be lowered. Overall Recommendation Ranking: After normalizing the above two scores, the final investment recommendation score is calculated. in , These are preset weighting coefficients, which typically prioritize ensuring enterprise quality. ; according to Sort the candidate companies in descending order and truncate the top few. Each company generates a targeted list of potential investment partners; For high-scoring companies on the list, extract publicly available information about the target companies and associate the contact paths.

Images