A management method and system for a micro portal platform

Abstract

The application discloses a kind of management method and system for micro portal platform, belong to building business management technical field, specifically include: based on the multi-source basic data of enterprise in building in the construction industry chain ecological map;Industry chain ecological map is extracted with upstream and downstream dependence relationship restoration to industry chain link, constructs industry chain structure model and identifies missing link;Filtering candidate enterprise set matched with missing link, obtain recruitment dynamic data and calculate space stress score;The number of already-registered enterprises that exist upstream and downstream dependence relationship with candidate enterprise is counted, calculate industry matching degree, and calculate comprehensive business value based on industry matching degree and space stress score;According to comprehensive business value descending order and carry out business reach task distribution.The application will enterprise dynamic site selection intention and building industry ecological demand joint quantitative evaluation, realize from passive fill in to active build ecology change, improve the precision of digital building business and industry ecological health degree.

Description

Technical Field

[0001] This invention relates to the field of building leasing management technology, specifically to a management method and system for a micro-portal platform. Background Technology

[0002] With the deepening of digital building construction, building operation and management has gradually evolved from the traditional property management model towards digitalization and intelligence. Currently, digital building management platforms integrate multi-source information such as enterprise registration data, property services, and policy information to achieve basic functions such as enterprise entry management, daily service response, and unified information dissemination. This has improved the standardization and service efficiency of building management to a certain extent, providing enterprises with a unified service entry point and information access channel.

[0003] However, existing digital building management platforms still have fundamental flaws in the core aspects of business attraction management. Current technologies generally treat businesses as static profiles, filtering them based on their current size and industry category, failing to capture the dynamic site selection needs arising from business expansion. Understandably, business relocation is a low-frequency, high-cost decision, and its decision-making window often first reveals itself through recruitment activities. When a company enters an expansion phase, it will be the first to post job openings, while existing office space gradually becomes saturated, creating "space pressure." This space pressure signal triggered by recruitment expansion is the most critical leading indicator for predicting a company's relocation intentions. Therefore, current technologies fail to translate the behavior of "the company is expanding its recruitment" into a quantitative assessment of "the company needs more space."

[0004] Meanwhile, existing technologies lack a joint assessment mechanism for the needs of the building's industrial ecosystem and the site selection intentions of enterprises. On the one hand, it is impossible to quantify the missing links in the industrial chain of the current tenant mix of the building; on the other hand, it is impossible to match and calculate the dynamic expansion needs of enterprises with the industrial positioning of the building. This leads to investment promotion decisions either focusing solely on "what industries the building lacks," blindly pursuing a closed industrial chain while ignoring the actual relocation intentions of enterprises; or focusing solely on "which enterprises are large-scale," simply screening existing enterprises while ignoring their space pressure. The result of this disconnect is that investment promotion reaches a large number of ineffective targets, enterprises with high relocation intentions are missed, and the optimization of the building's industrial ecosystem and investment promotion efficiency cannot be achieved simultaneously.

[0005] Against the backdrop of increasingly fierce competition in the regional building economy, if investment promotion management continues to rely on static profile screening and experience-based judgment, buildings will be in a long-term predicament of "being unable to retain attracted companies and being unable to find desired companies." High-quality companies will be forced to relocate due to space constraints, the building's tenant structure will be aging, and the industrial ecosystem will be unable to form a synergistic effect, ultimately restricting the continuous improvement of building asset value and regional industrial competitiveness. Summary of the Invention

[0006] The purpose of this invention is to provide a management method and system for micro-portal platforms, solving the following technical problems:

[0007] Existing building leasing management methods are unable to jointly and quantitatively assess the dynamic site selection intentions of enterprises with the needs of the building's industrial ecosystem, resulting in a serious disconnect between leasing decisions and the actual needs of enterprises.

[0008] The objective of this invention can be achieved through the following technical solutions: A management method for a micro-portal platform includes the following steps: S1, based on multi-source basic data of enterprises in the building, extracts the business scope characteristics and supply chain association characteristics of each enterprise, and constructs an industrial chain ecosystem map containing enterprise nodes and industry association edges; S2, extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map, construct an industrial chain structure model containing nodes of each industrial chain link based on the upstream and downstream dependencies between the links of the industrial chain, and identify the missing links of the current building in the industrial chain structure model. S3, filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data; S4, deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industrial matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industrial matching degree and the spatial pressure score. S5. Sort each candidate company in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and assign investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

[0009] As a further aspect of the present invention: the specific construction process of the industrial chain ecosystem map in S1 is as follows: Acquire multi-source basic data of each enterprise in the building, including business registration data, intellectual property data, and supply chain invoice data; The business scope text of each enterprise is extracted from the business registration data, and keyword extraction and industry classification coding mapping are performed on the business scope text to obtain the business scope characteristics of each enterprise. The patent classification number and trademark category of each enterprise are extracted from the intellectual property data, and the upstream and downstream transaction object information of each enterprise is extracted from the supply chain invoice data. The patent classification number, trademark category and upstream and downstream transaction object information are integrated to obtain the supply chain association characteristics of each enterprise. Using each enterprise as a node, the characteristics of each enterprise's business scope determine the identifier of the industrial chain link to which each enterprise belongs, and the characteristics of each enterprise's supply chain association determine the upstream and downstream transaction relationships between enterprises. An industrial association edge is established between two enterprises with upstream and downstream transaction relationships. All enterprise nodes and industrial association edges are aggregated to construct an industrial chain ecosystem map containing enterprise nodes and industrial association edges.

[0010] As a further aspect of the present invention: the specific construction process of the industrial chain structure model in S2 is as follows: Obtain all enterprise nodes in the industrial chain ecosystem map and the industrial chain link identifiers corresponding to each enterprise node. Deduplicate and aggregate the industrial chain link identifiers of all enterprise nodes to obtain multiple industrial chain link nodes. Obtain the industry association edges between each enterprise node in the industry chain ecosystem map. Based on the industry chain link identifiers corresponding to the two enterprise nodes connected by the industry association edge, determine the two industry chain link nodes associated with the industry association edge. Count the number of industry association edges between any two industry chain link nodes and use the number of industry association edges as the strength of the upstream and downstream dependency relationship between the two industry chain link nodes. Using each node in the industrial chain as a node and the strength of the upstream and downstream interdependence between any two nodes as edges, we construct an industrial chain structure model that includes the nodes and the edges of the upstream and downstream interdependence.

[0011] As a further aspect of the present invention: the specific process of identifying the missing link of the current building in the industrial chain structure model in step S2 is as follows: Obtain the industry chain link identifiers corresponding to all enterprise nodes in the industry chain ecosystem map, and deduplicate the industry chain link identifiers to obtain the set of industry chain links currently covered by the building; Obtain a preset regional industrial chain standard map, which includes multiple standard industrial chain nodes and upstream and downstream dependency edges between each standard industrial chain node. The set of industrial chain links currently covered by the building is compared with the standard industrial chain link nodes in the regional industrial chain standard map. Standard industrial chain link nodes that exist in the regional industrial chain standard map but are not included in the set of industrial chain links currently covered by the building are extracted as the initial missing links. For each initially selected missing link, standard industrial chain link nodes with upstream and downstream dependencies on the aforementioned regional industrial chain standard map are obtained. The number of nodes among the obtained standard industrial chain link nodes that belong to the industrial chain link set already covered by the current building is counted, and the number of nodes is used as the correlation between the initially selected missing link and the current building. The initially selected missing links with a correlation greater than a preset threshold are selected as key missing links, and these key missing links are considered as the missing links of the current building in the industrial chain structure model.

[0012] As a further aspect of the present invention: in S3, the specific process for generating the spatial pressure fraction is as follows: Obtain the job posting records of each company in the candidate company set within a preset time window. The job posting records include the job title, job description, and posting date. Obtain the number of social security contributors for each enterprise in the candidate enterprise set at the start and end times of the preset time window, calculate the difference between the number of social security contributors at the end time and the number of social security contributors at the start time, and obtain the net increase in personnel for each enterprise. The job title and job description are parsed to extract keywords from the job title and job description and work scenario keywords from the job description. The extracted keywords are matched with a preset keyword library for space-occupying jobs. Based on the matching results, each job is divided into space-occupying jobs and non-space-occupying jobs. The number of new space-occupying positions within the preset time window is counted to obtain the number of new space-occupying positions. Based on the net increase in personnel and the number of new space-occupying positions, the net increase in personnel for space-occupying positions is calculated. Obtain the office area data of each enterprise in the candidate enterprise set, take the sum of the number of social security contributors at the end time and the net expansion of personnel in the space-occupying positions as the expected number of personnel occupying space, calculate the quotient of the expected number of personnel occupying space and the office area, and obtain the space pressure score of each enterprise.

[0013] As a further aspect of the present invention: the specific construction process of the preset space-occupancy-type job keyword library is as follows: Obtain historical job posting sample data, which includes multiple job posting samples, each of which corresponds to a job title text and a job description text. The job title text and job description text are preprocessed, including word segmentation and stop word filtering, to obtain the keyword sequence for each job sample. Obtain the space occupancy attribute annotation results for each job sample. The space occupancy attribute annotation results include space-occupying type and space-non-occupying type. The space-occupying type indicates that the job requires a fixed workstation, and the space-non-occupying type indicates that the job is mainly field work and does not require a fixed workstation. Extract keyword sequences from space-occupying job samples with annotation results, and count the frequency of each keyword in space-occupying job samples; filter keywords with a frequency higher than a preset frequency threshold to construct an initial keyword library; Obtain the keyword sequence of job samples that are labeled as non-space-occupying, count the frequency of each keyword in the non-space-occupying job samples, and remove keywords in the initial keyword library that have a frequency higher than a preset interference threshold in the non-space-occupying job samples to obtain a keyword library for space-occupying jobs.

[0014] As a further aspect of the present invention: the specific calculation process for the industry matching degree in S4 is as follows: Obtain the upstream and downstream dependencies between nodes in each link of the industrial chain in the industrial chain structure model. The upstream and downstream dependencies include the direction and strength of the association between nodes in each link of the industrial chain. Obtain the set of stage identifiers of the enterprises currently located in the building, and the stage identifiers of each candidate enterprise in the candidate enterprise set; For each candidate enterprise, query the industry chain structure model for the industry chain link nodes that have an upstream dependency relationship with the link identifier of the candidate enterprise, obtain the set of upstream related link nodes of the candidate enterprise, and obtain the association strength between each upstream related link node and the link identifier of the candidate enterprise. The nodes that fall into the upstream related node set in the set of process identifiers of the enterprises that have settled in are counted. The association strength of each falling node is used as a weight, and the number of each falling node is counted as one. The product of the association strength of each falling node and the number of nodes is calculated, and all products are summed to obtain the upstream association value of the candidate enterprise. The upstream association value is used as the industry matching degree between the candidate enterprise and the current building.

[0015] As a further aspect of the present invention: the specific calculation process of the comprehensive investment value score in S4 is as follows: Obtain the industry matching degree and spatial pressure score of each candidate enterprise, and normalize the industry matching degree and spatial pressure score to obtain normalized industry matching degree and normalized spatial pressure score. The normalized industry matching degree and the normalized spatial pressure score are weighted and summed to obtain the comprehensive investment value score of each candidate enterprise.

[0016] The present invention also includes a management system for a micro-portal platform, for implementing the above-described management method for a micro-portal platform, comprising: The data acquisition module is used to extract the business scope characteristics and supply chain association characteristics of each enterprise based on multi-source basic data of enterprises in the building, and to construct an industrial chain ecosystem map that includes enterprise nodes and industry association edges. The missing link identification module is used to extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map. Based on the upstream and downstream dependencies between the links of the industrial chain, an industrial chain structure model containing the nodes of each link of the industrial chain is constructed, and the missing links of the current building in the industrial chain structure model are identified. The data analysis module is used to filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data. The value generation module is used to deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industry matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industry matching degree and the spatial pressure score. The task allocation module is used to sort the candidate companies in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and to allocate investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

[0017] The beneficial effects of this invention are: 1) Quantifying space pressure through dynamic recruitment data enables proactive prediction of enterprise relocation intentions. This invention breaks through the limitations of traditional investment promotion management's static understanding of enterprise needs. By collecting enterprise recruitment job records within a preset time window and combining text analysis of job titles and job descriptions, jobs are divided into two categories: space-occupying and space-free. Space-occupying jobs correspond to job categories requiring fixed workstations, and their increase directly reflects the actual workstation demand generated by the enterprise's business expansion. Space-free jobs correspond to job categories mainly involving fieldwork, and their increase does not constitute pressure on office space. Based on this, this invention further obtains the number of employees paying social security at the beginning and end of the time window, calculates the net increase in personnel, and combines it with the increase in space-occupying jobs to obtain the net expansion of personnel for space-occupying positions. Summing this net expansion with the current number of employees paying social security yields the expected number of personnel occupying space, which is then divided by the enterprise's office area to generate a space pressure score. This score accurately quantifies the space saturation level of the enterprise due to actual personnel expansion; a higher score indicates a more urgent relocation need. This mechanism transforms the behavioral signal of "expanding recruitment" into a quantifiable indicator of "needing more space," enabling investment promotion to shift from passively waiting for companies to release their needs to proactively capturing the window of opportunity for company expansion, significantly improving the timeliness and accuracy of investment promotion responses.

[0018] 2) By using an industry chain ecosystem map and calculating industry matching degree, this invention achieves accurate identification of missing links in the building's industry structure and targeted targeting of investment attraction goals. Based on multi-source basic data of enterprises within the building, this invention extracts business scope characteristics and supply chain association characteristics to construct an industry chain ecosystem map containing enterprise nodes and industry association edges. By extracting industry chain links and restoring upstream and downstream dependencies from the map, an industry chain structure model is formed with industry chain link nodes and the strength of upstream and downstream dependencies as edges. On this basis, a pre-set regional industry chain standard map is introduced for comparison to identify industry chain links existing in the regional industry chain but not currently covered by the building. Key missing links are then selected based on the strength of upstream and downstream dependencies of already covered links, clarifying the industry directions that the building needs to supplement. Furthermore, for the selected candidate enterprises, this invention queries the industry chain structure model for industry chain link nodes with upstream dependencies on the enterprise, counts the number of enterprises currently located in the building that fall into the set of associated nodes, and performs a weighted summation using the strength of upstream and downstream dependencies of each associated node as a weight to generate the industry matching degree between the candidate enterprise and the current building. This matching degree accurately reflects the strength of upstream and downstream supporting services that candidate companies can obtain after moving in, ensuring that the investment promotion goals are precisely aligned with the direction of optimizing the building's industrial ecosystem.

[0019] 3) By jointly assessing industry matching degree and spatial pressure, this invention achieves multi-objective comprehensive ranking and precise task allocation for investment attraction priorities. The invention normalizes the industry matching degree and spatial pressure scores of candidate companies to eliminate the difference in their dimensions, and then generates a comprehensive investment attraction value score through weighted summation. This weighting mechanism can flexibly adjust the weight coefficients according to actual operational strategies. For example, when a building is in the early stages of ecosystem construction, the focus can be on industry matching degree to prioritize filling key links in the industrial chain; when the building's ecosystem is relatively mature, the focus can be on spatial pressure scores to prioritize identifying companies with high relocation intentions. The comprehensive investment attraction value score simultaneously considers both the building's industrial ecosystem optimization needs and the companies' genuine relocation intentions, avoiding the one-sided problems of traditional investment attraction decisions that "only focus on missing industrial chains while ignoring company intentions" or "only focus on company size while ignoring ecosystem matching." Finally, the investment attraction priority sequence is output in descending order of the comprehensive investment attraction value score, guiding the investment attraction team to prioritize limited resources for companies with high building ecosystem matching degrees and genuine expansion needs, achieving a dual improvement in investment attraction efficiency and industrial ecosystem health, forming a complete investment attraction decision-making closed loop from ecosystem diagnosis and intention perception to precise outreach. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a schematic diagram of a management method for a micro-portal platform according to the present invention.

[0022] Figure 2 This is a schematic diagram of a management system structure for a micro portal platform according to the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 As shown, the present invention is a management method for a micro-portal platform, comprising the following steps: S1, based on multi-source basic data of enterprises in the building, extracts the business scope characteristics and supply chain association characteristics of each enterprise, and constructs an industrial chain ecosystem map containing enterprise nodes and industry association edges; S2, extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map, construct an industrial chain structure model containing nodes of each industrial chain link based on the upstream and downstream dependencies between the links of the industrial chain, and identify the missing links of the current building in the industrial chain structure model. S3, filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data; S4, deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industrial matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industrial matching degree and the spatial pressure score. S5. Sort each candidate company in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and assign investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

[0025] First, multi-source basic data of all resident enterprises within the building is acquired. This multi-source basic data includes business registration data, intellectual property data, and supply chain invoice data. The business scope text of each enterprise is extracted from the business registration data, and keyword extraction and industry classification coding are performed on the business scope text to obtain the business scope characteristics of each enterprise. These characteristics are used to determine the link in the industrial chain to which each enterprise belongs. Patent classification numbers and trademark categories of each enterprise are extracted from the intellectual property data, and upstream and downstream transaction information of each enterprise is extracted from the supply chain invoice data. The patent classification numbers, trademark categories, and upstream and downstream transaction information are integrated to obtain the supply chain association characteristics of each enterprise. Using each enterprise as a node, the industrial chain link identifier of each enterprise is determined based on its business scope characteristics, and the upstream and downstream transaction relationships between enterprises are determined based on their supply chain association characteristics. An industrial association edge is established between two enterprises with upstream and downstream transaction relationships. All enterprise nodes and industrial association edges are aggregated to construct an industrial chain ecosystem map containing enterprise nodes and industrial association edges. This map visually presents the industrial association structure among enterprises within the building, providing a data foundation for subsequent industrial chain analysis.

[0026] Based on the constructed industrial chain ecosystem map, an industrial chain structure model is built. All enterprise nodes in the industrial chain ecosystem map and their corresponding industrial chain link identifiers are obtained. The industrial chain link identifiers of all enterprise nodes are deduplicated and aggregated to obtain multiple industrial chain link nodes. The industrial association edges between enterprise nodes in the industrial chain ecosystem map are obtained. Based on the industrial chain link identifiers corresponding to the two enterprise nodes connected by the industrial association edge, the two industrial chain link nodes associated with that edge are determined. The number of industrial association edges between any two industrial chain link nodes is counted, and this number is used as the strength of the upstream and downstream dependency relationship between the two industrial chain link nodes. Using each industrial chain link node as a node and the strength of the upstream and downstream dependency relationship between any two industrial chain link nodes as edges, an industrial chain structure model containing industrial chain link nodes and upstream and downstream dependency relationship edges is constructed. This model abstracts enterprise-level micro-relationships into industrial chain link-level macro-relationships, facilitating the identification of the location and absence of buildings in the regional industrial chain.

[0027] To identify the missing links in the current building's industrial chain structure model, we first obtain the industrial chain link identifiers corresponding to all enterprise nodes in the industrial chain ecosystem map. After deduplication, we obtain the set of industrial chain links already covered by the current building. We then obtain a pre-defined regional industrial chain standard map, which contains multiple standard industrial chain link nodes and upstream / downstream dependency edges between these nodes, reflecting the ideal industrial chain structure in the regional industrial development plan. We compare the set of industrial chain links already covered by the current building with the standard industrial chain link nodes in the regional industrial chain standard map, extracting standard industrial chain link nodes that exist in the standard map but are not included in the covered set as initial missing links. For each initial missing link, we obtain standard industrial chain link nodes in the regional industrial chain standard map that have upstream / downstream dependency edges with that initial missing link. We count the number of nodes among the obtained standard industrial chain link nodes that belong to the set of industrial chain links already covered by the current building, and use this number as the correlation between the initial missing link and the current building. The correlation reflects the closeness of the missing link to the building's existing industrial ecosystem. In other words, the higher the correlation, the closer the missing link is to the building's existing business supply chain, and the stronger the synergistic effect that can be formed after introducing businesses in this link. Initially selected missing links with a correlation greater than a preset threshold are designated as key missing links, and these key missing links are output as the missing links in the current building's industry chain structure model.

[0028] After identifying the vacant positions in the building, companies matching these vacancies are selected from the regional enterprise database to form a candidate company set. For each company in the candidate set, its space pressure score needs to be calculated to quantify its relocation intention. This is achieved by obtaining the candidate companies' job posting records within a preset time window, including job titles, job descriptions, and posting dates. Simultaneously, the number of employees contributing social security at the start and end of the preset time window is obtained, and the difference between the start and end times is calculated to obtain the net increase in personnel. This value reflects the net change in the company's overall personnel size; a positive number represents expansion, and a negative number represents contraction. The job titles and job descriptions are parsed to extract keywords, which are then matched against a preset keyword database for space-occupying positions. Based on the matching results, each position is categorized into space-occupying and non-space-occupying positions. Space-occupying positions require fixed workstations, and their addition directly puts pressure on office space; non-space-occupying positions are primarily field-based, and their addition does not require fixed workstations. The number of new space-occupying positions added within a preset time window is statistically analyzed. Combined with the net increase in personnel, the net expansion of personnel for these positions is calculated. This net expansion excludes replacements due to employee departures, accurately reflecting the increased demand for workstations driven by business expansion. Office space data for candidate companies is obtained. The sum of the number of employees paying social security at the end of the period and the net expansion of personnel for space-occupying positions is used as the expected number of personnel occupying space. The quotient of this expected number to the office space is calculated to obtain a space pressure score. This score precisely quantifies the company's current space saturation level; a higher score indicates greater space pressure due to personnel expansion and a stronger willingness to relocate or expand.

[0029] The pre-defined space-occupancy-based keyword library for job positions is constructed as follows: Historical job posting sample data is obtained, each sample containing job title text and job description text. Preprocessing is performed on the text, including word segmentation and stop word filtering, to obtain keyword sequences. Space-occupancy attribute annotation results are obtained for each job sample, including space-occupying and non-space-occupying attributes. Keyword sequences are extracted from job samples labeled as space-occupying, and the frequency of each keyword is counted. Keywords with frequencies exceeding a preset threshold are selected to construct the initial keyword library. Further, keyword sequences are extracted from job samples labeled as non-space-occupying, and the frequency of each keyword is counted. Keywords from the initial keyword library whose frequency in non-space-occupying samples exceeds a preset interference threshold are removed, resulting in the final space-occupancy-based keyword library. This construction method ensures the accuracy and generalization ability of the keyword library.

[0030] After calculating the spatial pressure scores of candidate companies, the industry matching degree between each candidate company and the current building is further calculated. The industry chain structure model is deconstructed to obtain the upstream and downstream dependencies between nodes in each industry chain link. This dependency includes the direction of association (upstream or downstream) and the strength of association. The set of link identifiers for companies already located in the current building, as well as the link identifiers for each candidate company, are obtained. For each candidate company, the industry chain structure model is queried for the nodes in the industry chain that have an upstream dependency relationship with the candidate company's link identifier. This yields the set of upstream related links for the candidate company, and the strength of association between each upstream related link node and the candidate company is obtained. Upstream related links represent the supporting links that the candidate company needs in its production and operation. For example, a chip design company needs upstream EDA tool providers and IP core licensors. The system identifies nodes within the upstream related node set of existing resident companies, assigning each node its association strength as a weight and counting the number of nodes as one. It then calculates the product of the association strength and the number of nodes, summing all these products to obtain the upstream association value for the candidate company. This upstream association value is used as the industry matching degree between the candidate company and the current building. This matching degree reflects the strength of upstream and downstream support the candidate company can obtain after moving in. In other words, a higher matching degree indicates that existing resident companies can provide a more complete industrial chain support for the candidate company, resulting in lower operating costs and higher collaborative efficiency after the candidate company moves in.

[0031] The industry matching degree and spatial pressure score of each candidate enterprise are normalized to eliminate the difference in their dimensions and map them to the same numerical range. Normalization can be performed using maximum-minimum normalization or Z-score standardization. The normalized industry matching degree and spatial pressure score are then weighted and summed to obtain the comprehensive investment attraction value score of each candidate enterprise. The weighting coefficients can be dynamically adjusted according to the building operation strategy—when the building is in the early stage of ecosystem construction, the focus can be on industry matching degree to prioritize filling key links in the industrial chain; when the building ecosystem is relatively mature, the focus can be on spatial pressure score to prioritize identifying enterprises with a high willingness to relocate.

[0032] Finally, the candidate companies were ranked in descending order according to their comprehensive investment attraction value score, resulting in an investment attraction priority sequence. A higher comprehensive investment attraction value score indicates that the company is both highly compatible with the building's industrial ecosystem and has a genuine need for space expansion, making it the most worthy target for investment attraction. Based on this priority sequence, investment attraction tasks were assigned to the candidate companies, guiding the investment attraction team to prioritize high-value companies with limited resources, thereby achieving a dual improvement in investment attraction efficiency and the health of the industrial ecosystem.

[0033] In another preferred embodiment of the present invention, the specific construction process of the industrial chain ecosystem map in step S1 is as follows:The process involves acquiring multi-source basic data from each resident company within the building, including business registration data, intellectual property data, and supply chain invoice data. Taking a technology building as an example, the business scope text of each company is first extracted from the business registration data. For instance, if a resident company's business scope is registered as "computer software development, information technology consulting services, and data processing services," keyword extraction is performed to identify core business terms such as "computer software," "information technology," and "data processing." These terms are then mapped to the National Economic Industry Classification System to determine that the company belongs to the "software development" industry chain segment. Another company's business scope is registered as "computer hardware sales and system integration services," which is mapped to the "hardware distribution and system integration" segment. The reason why the industry chain segment to which a company belongs can be determined through the business scope text is that the business scope is the boundary of business activities declared by the company during business registration. Its text content directly reflects the company's main business direction and has an inherent correspondence with the National Economic Industry Classification System. By extracting patent classification numbers and trademark categories from intellectual property data, such as the aforementioned software development company holding multiple patents concentrated in the "G06F (Electrical Digital Data Processing)" field and its trademark categories covering "Class 9 (Scientific Instruments)" and "Class 42 (Scientific and Technological Services)," this information collectively reveals the company's professional attributes in terms of technological R&D direction and market brand positioning. By extracting upstream and downstream transaction information from supply chain invoice data, for example, the software development company's purchase invoices show that it has a long-term history of purchasing computing resources from a chip design company, while its sales invoices show that its software products are mainly sold to a system integrator. Integrating patent classification numbers, trademark categories, and upstream and downstream transaction information reveals the company's supply chain characteristics—patent classification numbers indicate that its technological capabilities are concentrated in the data processing field, trademark categories indicate that it targets the technology service market, and transaction partners clarify its position in the industry chain: upstream, it relies on chip design companies for computing power support, and downstream, it serves system integrators for final delivery. The reason why supply chain linkage characteristics can be constructed through patent classification numbers, trademark categories, and transaction object information is that patent classification numbers are internationally recognized technical field identifiers that can objectively reflect the company's technology research and development direction; trademark categories reflect the company's product or service positioning in the market; and upstream and downstream transaction objects are direct evidence of actual business transactions between companies. Together, these three constitute a three-dimensional portrait of the company's role in the industrial chain.Using each enterprise as a node, the industry chain segment to which each enterprise belongs is identified based on its business scope characteristics. Software development companies are labeled as belonging to the "software R&D" segment, chip design companies as belonging to the "chip design" segment, and system integrators as belonging to the "system integration" segment. The upstream and downstream transaction relationships between enterprises are determined based on their supply chain linkage characteristics—according to invoice records, if a chip design company and a software development company have transactions, they have an upstream supply relationship; if a software development company and a system integrator have transactions, they have a downstream supply relationship. An industry linkage edge is established between two enterprises with upstream and downstream transaction relationships, and the linkage direction and the strength of the linkage reflected by the transaction frequency are marked. Aggregating all enterprise nodes and industry linkage edges forms an industry chain ecosystem map containing enterprise nodes and industry linkage edges. This map visually presents the industry chain position of each enterprise within the building and the supply and demand cooperation relationships between enterprises.

[0034] The core purpose of constructing the industrial chain ecosystem map through the above methods is to transform the scattered and isolated enterprise data within the building into a structured industrial chain knowledge system. Specifically, the extraction of business scope characteristics allows for the accurate positioning of enterprises' main businesses within the national economic industry classification system, providing a standardized basis for subsequent industrial chain segmentation. The integration of supply chain association characteristics comprehensively depicts the role and collaborative relationships of enterprises within the industrial chain through three dimensions: patent technology direction, market brand positioning, and real transaction records. This multi-dimensional feature extraction and association construction allows industrial information originally hidden in business registration documents, intellectual property certificates, and transaction invoices to be mined and connected, forming an ecosystem map that intuitively reflects the building's industrial structure. This enables building operators to examine the industrial distribution of enterprises within the building from a global perspective for the first time, identify which industrial chain segments have formed clusters, which segments have gaps, and whether there is substantial supply and demand collaboration between enterprises. The establishment of this basic data asset provides a solid quantitative basis for subsequent identification of missing links in the industrial chain, screening of matching investment targets, and evaluation of the compatibility between candidate enterprises and the building ecosystem. It is the data cornerstone for the entire investment promotion management method to shift from "experience-driven" to "data-driven."

[0035] In another preferred embodiment of the present invention, the specific construction process of the industrial chain structure model in step S2 is as follows: First, we obtain all enterprise nodes and their corresponding industry chain segment identifiers from the industry chain ecosystem map constructed in the previous steps. In this map, enterprise nodes include chip design company A, chip design company B, software development company C, software development company D, and system integration company E. The industry chain segment identifiers for A and B are both "chip design," for C and D both "software development," and for E "system integration." We then perform deduplication and aggregation on the industry chain segment identifiers of all enterprise nodes, removing duplicates of "chip design," "software development," and "system integration," retaining only unique identifiers. This yields three industry chain segment nodes, corresponding to the segments of "chip design," "software development," and "system integration," respectively. The reason we can achieve the conversion from enterprise nodes to industry chain segment nodes through deduplication and aggregation is that industry chain analysis focuses on "segments" rather than "enterprises." Multiple enterprises engaged in the same business belong to the same industry chain segment; they are functionally homogeneous, and their contribution to the industry chain structure is reflected in the overall scale and activity of that segment, rather than individual differences. Next, we obtain the industry-related edges between the enterprise nodes in the industry chain ecosystem graph. For example, the graph contains four industry-related edges: chip design company A supplies chips to software development company C; chip design company B supplies chips to software development company D; software development company C supplies software to system integration company E; and software development company D also supplies software to system integration company E. Based on the industry chain segment identifiers corresponding to the two enterprise nodes connected by each industry-related edge, we determine the two industry chain segment nodes associated with that edge: the edge between A and C corresponds to the "chip design" and "software development" segments; the edge between B and D also corresponds to "chip design" and "software development"; the edge between C and E corresponds to "software development" and "system integration"; and the edge between D and E also corresponds to "software development" and "system integration". The number of industry-related edges between any two nodes in the industry chain is counted: there are two industry-related edges between "chip design" and "software development", so the upstream-downstream dependency strength between these two nodes is 2; there are also two industry-related edges between "software development" and "system integration", so the upstream-downstream dependency strength between these two nodes is also 2; there are no direct industry-related edges between "chip design" and "system integration", so the relationship strength is 0. The reason why the number of industry-related edges can be used as a measure of the upstream-downstream dependency strength is that the actual transactions or collaborations between enterprises directly reflect the supply and demand relationship in the industry chain. The more transactions, the more frequent the business interactions between the two links and the higher the degree of dependence. This relationship strength directly reflects the tightness between the links in the industry chain; the higher the strength, the closer the collaboration between the two links within the building and the more rigid their demand for each other.Finally, using each node in the industry chain as a node and the strength of the upstream-downstream dependency relationship between any two industry chain nodes as edges, an industry chain structure model is constructed that includes industry chain node links and their upstream-downstream dependency edges. In this model, there is an edge of strength 2 between "chip design" and "software development," another edge of strength 2 between "software development" and "system integration," and no edge between "chip design" and "system integration," thus forming a chain-like industry structure from "chip design" to "software development" and then to "system integration." This model abstracts the complex micro-network of enterprise-level connections into a concise macro-structural view at the link level, making the industry chain pattern, which was originally submerged in numerous enterprise nodes and transaction edges, clearly visible.

[0036] In another preferred embodiment of the present invention, the specific process of identifying the missing link of the current building in the industrial chain structure model in step S2 is as follows: First, obtain the industry chain segment identifiers corresponding to all enterprise nodes in the industry chain ecosystem map constructed in the previous steps. The companies located in this building include chip design company A, chip design company B, software development company C, and system integration company D, with corresponding industry chain segment identifiers of "chip design," "chip design," "software development," and "system integration," respectively. After deduplicating these identifiers, we obtain the set of industry chain segments currently covered by the building, namely {chip design, software development, system integration}. The reason we can obtain the covered segment set through deduplication is that industry chain analysis focuses on the existence of segments rather than the number of companies. As long as there is at least one company in a segment, the building is considered to have covered that segment. Next, a pre-defined regional industry chain standard map is obtained. This standard map is an ideal industry chain structure pre-constructed based on the regional industrial development plan. For example, if the region's industrial positioning is to develop a complete information technology industry chain of "chip design—software development—system integration—application services," then the standard map contains these four standard industry chain nodes and sets the upstream and downstream dependency relationships between each node. For example, "chip design" pointing to "software development" indicates an upstream supply relationship, "software development" pointing to "system integration" indicates an upstream supply relationship, and "system integration" pointing to "application services" indicates an upstream supply relationship. The current set of industry chain nodes covered by the building {chip design, software development, system integration} is compared with the standard industry chain nodes {chip design, software development, system integration, application services} in the regional industry chain standard map. The standard industry chain node that exists in the standard map but is not included in the current set of buildings covered by the building, namely the "application services" node, is extracted as the initial missing node. The reason why missing links can be identified through comparison is that the regional industrial chain standard map represents the ideal industrial structure that the region's industrial development aspires to achieve, while the set of links already covered by the building represents the current industrial reality of the building. The difference between the two is the building's weakness in the industrial ecosystem. For the initially selected missing link, "Application Services," standard industrial chain link nodes with upstream and downstream dependencies on "Application Services" are obtained from the regional industrial chain standard map. According to the standard map's setting, there is an upstream and downstream dependency between "System Integration" and "Application Services," meaning "System Integration" is an upstream link of "Application Services." Therefore, the obtained standard industrial chain link node is "System Integration." The number of nodes among the obtained standard industrial chain link nodes that belong to the current building's already covered industrial chain link set {chip design, software development, system integration} is counted. Since "System Integration" belongs to this set, the number of nodes is 1. This number of nodes is used as the correlation between the initially selected missing link, "Application Services," and the current building.The reason why the correlation degree can be calculated by counting the number of related nodes is that the richer the upstream and downstream dependencies between the missing link and the existing links in the building, the closer the missing link is to the existing industrial ecosystem of the building. If the upstream link of the missing link already exists in the building, the introduction of the missing link enterprise can obtain local supply chain support; if the downstream link of the missing link already exists in the building, its introduction can provide market channels for local enterprises; the more related nodes there are, the stronger the synergistic effect formed after the link is introduced. Assuming the preset threshold is set to 1, since the correlation degree of the missing link is 1, which is equal to the preset threshold, it is selected as a key missing link, and "application service" is output as the missing link of the current building in the industrial chain structure model. If there are multiple initially selected missing links, their respective correlation degrees are calculated, and only links with correlation degrees exceeding the preset threshold are selected as the final key missing links.

[0037] In another preferred embodiment of the present invention, the specific process for generating the spatial pressure fraction in step S3 is as follows: First, we obtain the company's job posting records within a preset time window, for example, the last three months. The company posted five job postings, including "Senior Software Engineer," "Front-end Development Engineer," "Regional Sales Manager," "After-sales Engineer," and "Administrative Specialist." Each posting includes the job title, job description, and posting date. Next, we obtain the number of employees contributing to social security at the beginning and end of the time window. For example, if the number of employees contributing to social security is 150 at the beginning and 165 at the end, the difference is calculated to obtain a net increase of 15 employees. The reason we can calculate the net increase in employees through the difference in the number of employees contributing to social security is that this is an authoritative record of the company's actual number of employees. Its changes directly reflect the net change in the company's workforce; a positive number represents expansion, and a negative number represents contraction. This data is more objective and timely than the number of employees disclosed by the company itself. Finally, we perform text parsing on the job titles and job descriptions of the five job postings, extracting keywords from the job titles and keywords related to the work scenarios in the job descriptions. For example, the job description for "Senior Software Engineer" includes keywords such as "participating in product development, coding, and tackling technical challenges," indicating a fixed workstation. The job description for "Regional Sales Manager" includes keywords such as "business trips to visit clients, market expansion, and attending industry conferences," indicating a predominantly travel-based work environment. The job description for "Administrative Specialist" includes keywords such as "office supplies management, front desk reception, and document archiving," also indicating a fixed workstation. The extracted keywords are matched against a pre-defined database of space-occupying job keywords. This database, trained using historical data, includes keywords for jobs requiring a fixed workstation, such as "R&D," "engineer," "development," "design," "specialist," and "assistant." Based on the matching results, "Senior Software Engineer," "Front-end Development Engineer," "After-sales Engineer," and "Administrative Specialist" match successfully and are classified as space-occupying jobs; "Regional Sales Manager" fails to match and is classified as a non-space-occupying job. The reason we can categorize job types using keyword matching is that the keywords in job titles and job descriptions are inherently related to the work scenarios of these positions. Jobs involving functions such as R&D, development, design, administration, and human resources typically require work at a fixed workstation, while positions involving functions such as sales, marketing, fieldwork, and consulting are primarily field-based and do not occupy a fixed workstation. Statistical analysis of a large number of historical job samples allows us to construct an accurate mapping relationship between keywords and job types. We counted the number of new "space-occupying" jobs added within a preset time window. Four out of the five job postings mentioned above were space-occupying jobs, therefore the number of new additions was 4.Based on the net increase of 15 employees and the addition of 4 space-occupying positions, the net expansion of space-occupying positions is calculated. This requires a comprehensive assessment of both: if the net increase in employees is positive and greater than the number of new space-occupying positions, it indicates strong overall expansion, and the net expansion of space-occupying positions can be taken as the number of new positions; if the net increase in employees is positive but less than the number of new space-occupying positions, it suggests potential job replacements or employee turnover, and the net increase in employees should be used as the upper limit; if the net increase in employees is negative, it indicates overall contraction, and even if space-occupying positions were created, it may be to replace departing employees, in which case the net expansion is counted as zero. In this case, the net increase of 15 employees is greater than the number of new space-occupying positions of 4, therefore the net expansion of space-occupying positions is 4. We obtained the company's office space data. Assuming the company currently leases 1000 square meters of office space, we calculated the expected number of employees occupying the space (169 people) by summing the current number of employees paying social security (165) at the end of the lease term and the net increase in staff for space-consuming positions (4). We then calculated the quotient of this expected number to the office space, resulting in a space pressure score of 0.169 people / square meter. This score reflects the company's current space occupancy density. A higher score indicates that more people need to be accommodated per unit area, the space is more crowded, and the company's need to relocate or expand due to staff expansion is more urgent.

[0038] The core value of calculating the space pressure score using the above method lies in transforming the vague behavioral signal of a company "currently hiring" into a quantifiable indicator that indicates "space saturation and the need for relocation," which can be used for investment decisions. Specifically, job postings are the most direct evidence of a company's expansion, but not all job postings imply space demand. Increased fieldwork positions do not occupy workstations, so text analysis is needed to differentiate job types, ensuring that only positions genuinely requiring fixed workstations are included in the space pressure calculation. Changes in the number of employees contributing to social security reflect the true changes in the company's workforce. Combining this with the number of new positions eliminates hiring spuriously due to employee departures, ensuring that net expansion reflects genuine business expansion. Office space data provides a benchmark for current space capacity. By using the ratio of expected staff to office space, absolute staff expansion is transformed into relative space saturation. The value of this quantifiable indicator lies in its ability to allow building operators to quickly identify companies with genuine relocation needs—those with high space pressure scores—who, understandably, have the strongest desire to relocate because their existing space cannot accommodate their expanded teams. This allows investment outreach to shift from a "broad-based information push" to "proactive communication targeting high-interest companies," enabling the allocation of limited human and resource resources to the most likely targets for success. Simultaneously, this score provides a quantitative basis for the subsequent calculation of the overall investment value score, reflecting the "real needs of enterprises." Together with industry matching, it forms a dual evaluation standard for investment prioritization, ensuring that the ultimately recommended investment targets are both needed by the building's industrial ecosystem and urgently required by the companies themselves, thereby maximizing the success rate of investment attraction and the efficiency of building ecosystem optimization.

[0039] In another preferred embodiment of the present invention, the specific construction process of the preset space-occupancy-type job keyword library is as follows: Obtain historical job posting sample data, which includes multiple job posting samples, each of which corresponds to a job title text and a job description text. The job title text and job description text are preprocessed, including word segmentation and stop word filtering, to obtain the keyword sequence for each job sample. Obtain the space occupancy attribute annotation results for each job sample. The space occupancy attribute annotation results include space-occupying type and space-non-occupying type. The space-occupying type indicates that the job requires a fixed workstation, and the space-non-occupying type indicates that the job is mainly field work and does not require a fixed workstation. Extract keyword sequences from space-occupying job samples with annotation results, and count the frequency of each keyword in space-occupying job samples; filter keywords with a frequency higher than a preset frequency threshold to construct an initial keyword library; Obtain the keyword sequence of job samples that are labeled as non-space-occupying, count the frequency of each keyword in the non-space-occupying job samples, and remove keywords in the initial keyword library that have a frequency higher than a preset interference threshold in the non-space-occupying job samples to obtain a keyword library for space-occupying jobs.

[0040] In another preferred embodiment of the present invention, the specific calculation process of the industry matching degree in step S4 is as follows: The building currently houses companies A (chip design), B (chip design), C (software development), and D (system integration), whose respective industry segments are "chip design," "chip design," "software development," and "system integration." A candidate company, E (application services), is also included, with its industry segment segment identified as "application services." The goal is to calculate the industry compatibility between this candidate company and the building. First, the upstream and downstream dependencies between nodes in the industry chain structure model are obtained. In this model, there is an upstream / downstream dependency edge between "chip design" and "software development," with the direction from chip design to software development, indicating that chip design is an upstream segment of software development, with a correlation strength of 2 (representing two industry-related edges). Similarly, there is an upstream / downstream dependency edge between "software development" and "system integration," with the direction from software development to system integration, also with a correlation strength of 2. Finally, there is an upstream / downstream dependency edge between "system integration" and "application services," with the direction from system integration to application services, with a correlation strength of 1 (representing one industry-related edge). The reason for needing to obtain upstream and downstream dependencies and their strength beforehand is that these relationships are fundamental data for measuring the degree of dependence between links in the industry chain. A stronger correlation indicates more frequent business interactions and closer collaboration between the two links, which has a significant weighting effect on subsequent matching degree calculations. Next, we obtain the set of links to which the currently resident companies belong, namely {chip design, software development, system integration}, and the link to which candidate company E belongs, "application services". For candidate company E, we query the industry chain structure model for links that have an upstream dependency relationship with "application services," i.e., we query which links are upstream of "application services". Based on the dependency relationships in the model, system integration points to application services; therefore, "system integration" is an upstream link of "application services". We also query whether there are other links that also point to application services. Assuming that "software development" does not directly point to "application services" in the model, we obtain the set of upstream related links for this candidate company as {system integration}, and the correlation strength between this upstream related link and the candidate company is set to 1. The reason for only querying upstream dependencies rather than downstream relationships is that industry matching focuses on whether candidate companies can obtain local supply chain support after moving in. Understandably, the more upstream links are located within the building and the stronger the connection, the more raw materials, components, or technical services the candidate company can obtain from within the building, thereby reducing procurement costs and shortening supply cycles. This local support capability is a crucial factor in company location decisions. The system identifies nodes within the upstream related link node set {system integration} of the already-located companies' segment identifier set {chip design, software development, system integration}. Since "system integration" belongs to the already-located set, the node it falls into is "system integration".The association strength of a node is used as the weight, with the association strength of the node being 1. The number of nodes falling into the node is also counted as 1. The product of the association strength of the falling node and the number of nodes is calculated as 1 multiplied by 1, which equals 1. Since only one node falls into the upstream association node set, the upstream association value of the candidate enterprise is summed to obtain 1. This upstream association value is used as the industry matching degree between candidate enterprise E and the current building. Suppose there is another candidate enterprise F, a hardware supply company, whose link is identified as "hardware supply". The upstream association node set of F is queried in the industry chain structure model. Assuming that "chip design" is one of its upstream links, and "chip design" is included in the already established set, the association strength of "chip design" will be counted when calculating the matching degree. If the association strength of "chip design" is 2, the matching degree is higher. In this way, the industry matching degree of each candidate enterprise reflects the local supporting capabilities of the building's existing industry ecosystem after its establishment.

[0041] The core value of calculating industry matching degree using the above method lies in transforming the synergistic potential between candidate companies and the building's existing industrial ecosystem into quantifiable and comparable numerical indicators. Specifically, by querying the upstream related nodes of candidate companies, the core question is "what kind of local support does the candidate company need?" Any company, when choosing a location, wants to be close to its suppliers to reduce logistics costs, shorten response time, and obtain technological synergy. The more upstream companies already located in the building that are related to the candidate companies, and the stronger the connection, the more comprehensive the supporting facilities the building can provide for the candidate companies, resulting in higher operational efficiency and lower costs after the candidate companies move in. This method of calculating matching degree transforms the abstract concept of "industrial ecosystem synergy" into an objective assessment based on actual transaction data, avoiding the subjectivity and uncertainty of judgments based on experience. Industry matching degree and spatial pressure score together constitute the two pillars of the comprehensive investment attraction value score. The spatial pressure score answers the question of "whether the company has the intention to relocate," while the industry matching degree answers the question of "whether the company is suitable to move into this building." Both are indispensable. Only when a candidate company has both a genuine need for expansion (high space pressure score) and a high degree of alignment with the building's existing industrial ecosystem (high industry matching degree) can it become the optimal investment attraction target. Therefore, the accurate calculation of industry matching degree ensures that investment attraction decisions are not blindly pursuing "filling vacant areas," but rather serving the continuous optimization and structural improvement of the building's industrial ecosystem. This allows each new resident company to create synergies with existing companies, enhancing the building's overall industrial stickiness, risk resistance, and long-term competitiveness.

[0042] In another preferred embodiment of the present invention, the specific calculation process of the comprehensive investment value score in step S4 is as follows: Obtain the industry matching degree and spatial pressure score of each candidate enterprise, and normalize the industry matching degree and spatial pressure score to obtain normalized industry matching degree and normalized spatial pressure score. The normalized industry matching degree and the normalized spatial pressure score are weighted and summed to obtain the comprehensive investment value score of each candidate enterprise.

[0043] Understandably, this approach breaks away from the traditional one-dimensional evaluation model of investment attraction decisions, which focuses solely on "what the building needs" or "what the company wants," and instead constructs a comprehensive evaluation framework that considers both supply and demand. Companies with high overall investment attraction value scores are not only key components that the building's industrial ecosystem needs to fill, but also have a strong need to relocate or expand due to space saturation. These companies are the ideal targets for building investment attraction because introducing them can optimize the building's industrial ecosystem structure and increase the success rate of investment attraction. Ultimately, by sorting investment attraction in descending order of overall investment attraction value scores to form an investment attraction priority sequence, building operators can prioritize the allocation of limited investment attraction resources to the most valuable candidate companies, maximizing investment attraction efficiency and optimizing the industrial ecosystem. This is the ultimate goal of this solution's transformation from "passively filling vacancies" to "proactively building an ecosystem."

[0044] See Figure 2 The present invention also includes a management system for a micro-portal platform, for implementing the above-described management method for a micro-portal platform, comprising: The data acquisition module is used to extract the business scope characteristics and supply chain association characteristics of each enterprise based on multi-source basic data of enterprises in the building, and to construct an industrial chain ecosystem map that includes enterprise nodes and industry association edges. The missing link identification module is used to extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map. Based on the upstream and downstream dependencies between the links of the industrial chain, an industrial chain structure model containing the nodes of each link of the industrial chain is constructed, and the missing links of the current building in the industrial chain structure model are identified. The data analysis module is used to filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data. The value generation module is used to deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industry matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industry matching degree and the spatial pressure score. The task allocation module is used to sort the candidate companies in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and to allocate investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

[0045] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A management method for a micro-portal platform, characterized in that, Includes the following steps: S1, based on multi-source basic data of enterprises in the building, extracts the business scope characteristics and supply chain association characteristics of each enterprise, and constructs an industrial chain ecosystem map containing enterprise nodes and industry association edges; S2, extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map, construct an industrial chain structure model containing nodes of each industrial chain link based on the upstream and downstream dependencies between the links of the industrial chain, and identify the missing links of the current building in the industrial chain structure model. S3, filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data; S4, deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industrial matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industrial matching degree and the spatial pressure score. S5. Sort each candidate company in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and assign investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

2. The management method for a micro-portal platform according to claim 1, characterized in that, In S1, the specific construction process of the industrial chain ecosystem map is as follows: Acquire multi-source basic data of each enterprise in the building, including business registration data, intellectual property data, and supply chain invoice data; The business scope text of each enterprise is extracted from the business registration data, and keyword extraction and industry classification coding mapping are performed on the business scope text to obtain the business scope characteristics of each enterprise. The patent classification number and trademark category of each enterprise are extracted from the intellectual property data, and the upstream and downstream transaction object information of each enterprise is extracted from the supply chain invoice data. The patent classification number, trademark category and upstream and downstream transaction object information are integrated to obtain the supply chain association characteristics of each enterprise. Using each enterprise as a node, the characteristics of each enterprise's business scope determine the identifier of the industrial chain link to which each enterprise belongs, and the characteristics of each enterprise's supply chain association determine the upstream and downstream transaction relationships between enterprises. An industrial association edge is established between two enterprises with upstream and downstream transaction relationships. All enterprise nodes and industrial association edges are aggregated to construct an industrial chain ecosystem map containing enterprise nodes and industrial association edges.

3. The management method for a micro-portal platform according to claim 1, characterized in that, In S2, the specific construction process of the industrial chain structure model is as follows: Obtain all enterprise nodes in the industrial chain ecosystem map and the industrial chain link identifiers corresponding to each enterprise node. Deduplicate and aggregate the industrial chain link identifiers of all enterprise nodes to obtain multiple industrial chain link nodes. Obtain the industry association edges between each enterprise node in the industry chain ecosystem map. Based on the industry chain link identifiers corresponding to the two enterprise nodes connected by the industry association edge, determine the two industry chain link nodes associated with the industry association edge. Count the number of industry association edges between any two industry chain link nodes and use the number of industry association edges as the strength of the upstream and downstream dependency relationship between the two industry chain link nodes. Using each node in the industrial chain as a node and the strength of the upstream and downstream interdependence between any two nodes as edges, we construct an industrial chain structure model that includes the nodes and the edges of the upstream and downstream interdependence.

4. The management method for a micro-portal platform according to claim 1, characterized in that, In step S2, the specific process of identifying the missing link of the current building in the industry chain structure model is as follows: Obtain the industry chain link identifiers corresponding to all enterprise nodes in the industry chain ecosystem map, and deduplicate the industry chain link identifiers to obtain the set of industry chain links currently covered by the building; Obtain a preset regional industrial chain standard map, which includes multiple standard industrial chain nodes and upstream and downstream dependency edges between each standard industrial chain node. The set of industrial chain links currently covered by the building is compared with the standard industrial chain link nodes in the regional industrial chain standard map. Standard industrial chain link nodes that exist in the regional industrial chain standard map but are not included in the set of industrial chain links currently covered by the building are extracted as the initial missing links. For each initially selected missing link, standard industrial chain link nodes with upstream and downstream dependencies on the aforementioned regional industrial chain standard map are obtained. The number of nodes among the obtained standard industrial chain link nodes that belong to the industrial chain link set already covered by the current building is counted, and the number of nodes is used as the correlation between the initially selected missing link and the current building. The initially selected missing links with a correlation greater than a preset threshold are selected as key missing links, and these key missing links are considered as the missing links of the current building in the industrial chain structure model.

5. A management method for a micro-portal platform according to claim 1, characterized in that, In S3, the specific process for generating the spatial pressure fraction is as follows: Obtain the job posting records of each company in the candidate company set within a preset time window. The job posting records include the job title, job description, and posting date. Obtain the number of social security contributors for each enterprise in the candidate enterprise set at the start and end times of the preset time window, calculate the difference between the number of social security contributors at the end time and the number of social security contributors at the start time, and obtain the net increase in personnel for each enterprise. The job title and job description are parsed to extract keywords from the job title and job description and work scenario keywords from the job description. The extracted keywords are matched with a preset keyword library for space-occupying jobs. Based on the matching results, each job is divided into space-occupying jobs and non-space-occupying jobs. The number of new space-occupying positions within the preset time window is counted to obtain the number of new space-occupying positions. Based on the net increase in personnel and the number of new space-occupying positions, the net increase in personnel for space-occupying positions is calculated. Obtain the office area data of each enterprise in the candidate enterprise set, take the sum of the number of social security contributors at the end time and the net expansion of personnel in the space-occupying positions as the expected number of personnel occupying space, calculate the quotient of the expected number of personnel occupying space and the office area, and obtain the space pressure score of each enterprise.

6. A management method for a micro-portal platform according to claim 5, characterized in that, The specific construction process of the preset space-occupancy-type job keyword library is as follows: Obtain historical job posting sample data, which includes multiple job posting samples, each of which corresponds to a job title text and a job description text. The job title text and job description text are preprocessed, including word segmentation and stop word filtering, to obtain the keyword sequence for each job sample. Obtain the space occupancy attribute annotation results for each job sample. The space occupancy attribute annotation results include space-occupying type and space-non-occupying type. The space-occupying type indicates that the job requires a fixed workstation, and the space-non-occupying type indicates that the job is mainly field work and does not require a fixed workstation. Extract keyword sequences from space-occupying job samples with annotation results, and count the frequency of each keyword in space-occupying job samples; filter keywords with a frequency higher than a preset frequency threshold to construct an initial keyword library; Obtain the keyword sequence of job samples that are labeled as non-space-occupying, count the frequency of each keyword in the non-space-occupying job samples, and remove keywords in the initial keyword library that have a frequency higher than a preset interference threshold in the non-space-occupying job samples to obtain a keyword library for space-occupying jobs.

7. A management method for a micro-portal platform according to claim 1, characterized in that, In S4, the specific calculation process for the industry matching degree is as follows: Obtain the upstream and downstream dependencies between nodes in each link of the industrial chain in the industrial chain structure model. The upstream and downstream dependencies include the direction and strength of the association between nodes in each link of the industrial chain. Obtain the set of stage identifiers of the enterprises currently located in the building, and the stage identifiers of each candidate enterprise in the candidate enterprise set; For each candidate enterprise, query the industry chain structure model for the industry chain link nodes that have an upstream dependency relationship with the link identifier of the candidate enterprise, obtain the set of upstream related link nodes of the candidate enterprise, and obtain the association strength between each upstream related link node and the link identifier of the candidate enterprise. The nodes that fall into the upstream related node set in the set of process identifiers of the enterprises that have settled in are counted. The association strength of each falling node is used as a weight, and the number of each falling node is counted as one. The product of the association strength of each falling node and the number of nodes is calculated, and all products are summed to obtain the upstream association value of the candidate enterprise. The upstream association value is used as the industry matching degree between the candidate enterprise and the current building.

8. A management method for a micro-portal platform according to claim 1, characterized in that, In S4, the specific calculation process for the comprehensive investment value score is as follows: Obtain the industry matching degree and spatial pressure score of each candidate enterprise, and normalize the industry matching degree and spatial pressure score to obtain normalized industry matching degree and normalized spatial pressure score. The normalized industry matching degree and the normalized spatial pressure score are weighted and summed to obtain the comprehensive investment value score of each candidate enterprise.

9. A management system for a micro-portal platform, used to implement the management method for a micro-portal platform as described in any one of claims 1-8, characterized in that, include: The data acquisition module is used to extract the business scope characteristics and supply chain association characteristics of each enterprise based on multi-source basic data of enterprises in the building, and to construct an industrial chain ecosystem map that includes enterprise nodes and industry association edges. The missing link identification module is used to extract the links of the industrial chain and restore the upstream and downstream dependencies of the industrial chain ecosystem map. Based on the upstream and downstream dependencies between the links of the industrial chain, an industrial chain structure model containing the nodes of each link of the industrial chain is constructed, and the missing links of the current building in the industrial chain structure model are identified. The data analysis module is used to filter the set of candidate companies that match the missing link, obtain the recruitment dynamic data of each company in the set of candidate companies, and calculate the spatial pressure score of each company based on the recruitment dynamic data. The value generation module is used to deconstruct the industrial chain structure model, obtain the segment identifier of the enterprises already in the building, count the number of enterprises that have upstream and downstream dependencies with each candidate enterprise, calculate the industry matching degree between each candidate enterprise and the current building, and calculate the comprehensive investment attraction value score of each candidate enterprise based on the industry matching degree and the spatial pressure score. The task allocation module is used to sort the candidate companies in descending order according to the comprehensive investment promotion value score to obtain an investment promotion priority sequence, and to allocate investment promotion outreach tasks to each candidate company according to the investment promotion priority sequence.

Images