Intelligent industrial chain construction method and system based on multi-dimensional qualification determination

Abstract

The invention discloses an intelligent industrial chain construction method and system based on multi-dimensional qualification affirmation, and the method comprises the steps: constructing a multi-dimensional qualification evaluation model based on industrial characteristic data; obtaining the preprocessed multi-dimensional qualification data of the industry chain related enterprises; analyzing the multi-dimensional qualification data by using a multi-dimensional qualification evaluation model to obtain a multi-dimensional quantitative index of the qualification data of the industry chain related enterprises; based on the multi-dimensional quantitative indexes, a machine learning algorithm is adopted to carry out classification to obtain enterprise category distribution, and comprehensive scoring and sorting are carried out on industry chain related enterprises according to set weights to obtain a sorting result; interaction data, transaction intensity data and resource flow data among related enterprises of the industry chain are acquired; and constructing an intelligent industrial chain network diagram according to the data obtained in the steps. The method is helpful for understanding an industrial chain structure, identifying key enterprises and evaluating potential risks.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial chain management, and in particular to a method and system for constructing an intelligent industrial chain based on multi-dimensional qualification certification. Background Art

[0002] With the rapid development of the global economy and the continuous optimization of the industrial structure, industrial chain management plays an increasingly important role in corporate operations and regional economic development. However, the traditional industrial chain construction method mainly relies on manual experience and single-dimensional enterprise evaluation, and there are many problems that need to be solved. First, the evaluation dimension of the traditional method is single, often focusing only on single factors such as the company's financial status or market share, ignoring the company's comprehensive capabilities in technological innovation, environmental protection, supply chain management, etc., and therefore cannot fully reflect the company's true strength.

[0003] Secondly, the difficulty in obtaining data is also a major problem of traditional methods. Enterprise-related data are scattered in different systems and platforms, the data format is not unified, and it is difficult to obtain and integrate, which seriously affects the accuracy and timeliness of the evaluation. In addition, the simplification of traditional evaluation models also restricts the accuracy of the evaluation. Usually, traditional models use simple weighted average or linear regression methods, which are difficult to capture the complex nonlinear relationships in enterprise qualification evaluation, resulting in inaccurate evaluation results.

[0004] In addition, the traditional method lacks dynamism and is difficult to reflect the interactive relationship between enterprises and changes in the market environment in a timely manner, which makes it unable to provide effective support for the dynamic optimization of the industrial chain. At the same time, the traditional method of industrial chain construction also lacks systematicity, often focusing only on a single enterprise or a local industrial chain, lacking systematic analysis and optimization of the entire industrial chain, making it difficult to achieve overall coordination and efficiency improvement of the industrial chain.

[0005] Finally, the traditional methods have a low level of intelligence and mainly rely on manual judgment and simple mathematical models. They cannot fully utilize the advantages of big data and artificial intelligence technologies, resulting in a low level of intelligence.

[0006] Therefore, there is an urgent need for a new method that can comprehensively evaluate enterprise qualifications and intelligently construct industrial chains in order to improve the overall efficiency and competitiveness of the industrial chain. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide an intelligent industrial chain construction method and system based on multi-dimensional qualification certification, aiming to achieve systematic construction and optimization of the industrial chain by comprehensively evaluating enterprise qualifications, intelligently analyzing industry characteristics and enterprise interactions.

[0008] To achieve the above objectives, in the first aspect, a method for constructing an intelligent industrial chain based on multi-dimensional qualification certification is provided, which comprises the following steps:

[0009] Build a multi-dimensional qualification assessment model based on industry characteristic data;

[0010] Obtain pre-processed multi-dimensional qualification data of related enterprises in the industrial chain;

[0011] Analyze the multi-dimensional qualification data using the multi-dimensional qualification evaluation model to obtain multi-dimensional quantitative indicators of the qualification data of enterprises related to the industrial chain;

[0012] Based on the multi-dimensional quantitative indicators of the qualification data, a machine learning algorithm is used to classify the enterprises related to the industrial chain to obtain the enterprise category distribution, and the enterprises related to the industrial chain are comprehensively scored and ranked according to the set weights to obtain the ranking results;

[0013] Obtain interaction data, transaction intensity data, and resource flow data between related enterprises in the industrial chain;

[0014] An intelligent industrial chain network diagram is constructed based on the multi-dimensional quantitative indicators, the enterprise category distribution, the ranking results, the industry characteristic data, the interaction data, the transaction intensity data and the resource flow data.

[0015] Second, a smart industry chain construction system based on multi-dimensional qualification certification is provided, which includes:

[0016] The qualification evaluation model construction module is used to build a multi-dimensional qualification evaluation model based on industry characteristic data;

[0017] The data acquisition module is used to obtain the pre-processed multi-dimensional qualification data of related enterprises in the industrial chain;

[0018] A qualification analysis module, used to analyze the multi-dimensional qualification data using the multi-dimensional qualification evaluation model to obtain multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain;

[0019] An enterprise classification and ranking module is used to classify the enterprises related to the industrial chain based on the multi-dimensional quantitative indicators of the qualification data, adopt a machine learning algorithm to obtain the enterprise category distribution, and comprehensively score and rank the enterprises related to the industrial chain according to the set weights to obtain the ranking results;

[0020] Interaction data acquisition module, used to obtain interaction data, transaction intensity data and resource flow data between related enterprises in the industrial chain;

[0021] A network diagram construction module is used to construct an intelligent industrial chain network diagram based on the multi-dimensional quantitative indicators, the enterprise category distribution, the sorting results, the industry characteristic data, the interaction data, the transaction intensity data and the resource flow data.

[0022] In a third aspect, an electronic device is provided, comprising:

[0023] one or more processors;

[0024] a storage device for storing one or more programs,

[0025] When the one or more programs are executed by the one or more processors, the one or more processors implement any method as described in the first aspect.

[0026] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method described in any one of the first aspects is implemented.

[0027] The above technical solution has the following beneficial technical effects:

[0028] By building a multi-dimensional qualification evaluation model and obtaining multi-dimensional qualification data, this method can comprehensively evaluate the various capabilities and characteristics of enterprises in the industrial chain. This comprehensive evaluation not only takes into account traditional financial indicators, but also includes multiple dimensions such as technological innovation, environmental compliance, and supply chain management, thus providing a more three-dimensional and accurate corporate portrait.

[0029] By classifying and ranking enterprises based on multi-dimensional quantitative indicators, enterprises in the industrial chain can be divided and compared more accurately. This classification and ranking result helps to identify industry leaders, discover potential enterprises, and understand the distribution of different types of enterprises in the industrial chain.

[0030] By constructing an intelligent industrial chain network diagram, this method presents the complex relationships between enterprises in an intuitive way. This visualization not only includes the attribute information of the enterprise, but also shows the interactive relationships between enterprises, such as supply relationships, cooperation relationships, and competition relationships, making the structure and dynamics of the industrial chain clear at a glance.

[0031] Since this method can continuously acquire and update enterprise data and interactive information, it can realize dynamic monitoring of the industrial chain. This dynamic monitoring capability helps to timely discover changes and potential risks in the industrial chain and provide early warnings for decision makers.

[0032] By comprehensively analyzing the multi-dimensional indicators, category distribution, ranking results and interactive data of enterprises, this method provides strong decision-making support for industrial chain optimization. Decision makers can formulate more accurate industrial policies based on this information, optimize resource allocation and promote the coordinated development of the industrial chain.

[0033] By comprehensively understanding the industrial chain structure and enterprise characteristics, this method helps identify weak links and key nodes in the industrial chain. This insight enables decision makers to take targeted measures, such as cultivating key enterprises and strengthening weak links, thereby enhancing the resilience and risk resistance of the entire industrial chain.

[0034] By showing the technological dependencies and cooperation relationships between enterprises, this method provides an important reference for innovation collaboration in the industrial chain. Enterprises can find potential partners based on this information, and the government can formulate more targeted industry-university-research cooperation policies, thereby promoting the innovation capability of the entire industrial chain.

[0035] Based on the multi-dimensional evaluation results of enterprises and the network structure of the industrial chain, this method provides a scientific basis for the optimal allocation of resources. Whether it is the government's industrial support funds or the flow of resources between enterprises, they can be allocated more reasonably and efficiently based on this information. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0037] Figure 1 It is a flow chart of a method for constructing an intelligent industrial chain based on multi-dimensional qualification certification according to an embodiment of the present invention;

[0038] Figure 2 is a specific flow chart of step S1 of an embodiment of the present invention;

[0039] Figure 3 is a specific flow chart of step S2 of an embodiment of the present invention;

[0040] Figure 4 is a specific flow chart of step S3 of an embodiment of the present invention;

[0041] Figure 5 is a specific flow chart of step S4 of an embodiment of the present invention;

[0042] Figure 6 is a specific flow chart of step S6 of an embodiment of the present invention;

[0043] Figure 7 is a specific flow chart of step S7 of an embodiment of the present invention;

[0044] Figure 8It is a functional block diagram of an intelligent industrial chain construction system based on multi-dimensional qualification certification according to an embodiment of the present invention;

[0045] Fig. 9 It is a schematic diagram of the structure of a computer system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0047] Example 1

[0048] like Figure 1 As shown, this embodiment provides a method for constructing an intelligent industrial chain based on multi-dimensional qualification certification, including the following steps: S1: construct a multi-dimensional qualification evaluation model based on industry characteristic data; S2: obtain multi-dimensional qualification data of pre-processed enterprises related to the industrial chain; S3: use the multi-dimensional qualification evaluation model to analyze the multi-dimensional qualification data to obtain multi-dimensional quantitative indicators of the qualification data of enterprises related to the industrial chain; S4: based on the multi-dimensional quantitative indicators of the qualification data, use a machine learning algorithm to classify enterprises related to the industrial chain to obtain enterprise category distribution, and comprehensively score and sort the enterprises related to the industrial chain according to the set weights to obtain sorting results; S5: obtain interaction data, transaction intensity data and resource flow data between enterprises related to the industrial chain; S6: construct an intelligent industrial chain network diagram based on multi-dimensional quantitative indicators, enterprise category distribution, sorting results, industry characteristic data, interaction data, transaction intensity data and resource flow data. The following will describe each step of the method in detail.

[0049] Step S1: Construct a multi-dimensional qualification evaluation model. Industry characteristic data refers to a data set that can reflect the unique properties of a specific industry, such as production efficiency, supply chain complexity, etc. in the manufacturing industry, or user growth rate, platform activity, etc. in the Internet industry. The multi-dimensional qualification evaluation model is an evaluation system that comprehensively considers multiple factors, including but not limited to financial dimensions (such as operating income, profit margin), technical dimensions (such as R&D investment ratio, number of patents), market dimensions (such as market share, brand value), management dimensions (such as management team experience, organizational structure efficiency) and innovation dimensions (such as new product development speed, innovation results conversion rate). During the construction process, the specific indicators and weights of each dimension need to be adjusted according to the characteristics of different industries. For example, high-tech industries may pay more attention to technology and innovation dimensions, while traditional manufacturing industries may pay more attention to financial and market dimensions.

[0050] Step S2: Obtain the preprocessed enterprise qualification data. The preprocessing in this step refers to the process of cleaning, standardizing and normalizing the original data. Specifically, first collect the original data of related enterprises in the industrial chain through various channels, including but not limited to public financial reports, patent databases, market research reports, government public information, etc. Then, these original data are cleaned (remove erroneous, duplicate or irrelevant data), standardized (convert data from different sources and formats into a unified standard format) and normalized (convert data of different dimensions to the same scale, usually the [0,1] interval). After these preprocessing steps, multi-dimensional qualification data that can be directly used for subsequent analysis is obtained.

[0051] Step S3: Analyze and obtain multi-dimensional quantitative indicators. In this step, the pre-processed multi-dimensional qualification data is input into the evaluation model constructed in step S1. Through data analysis and calculation, the quantitative indicators of each enterprise in each dimension are obtained. For example, the quantitative indicators of each dimension of a certain enterprise may be as follows: 82 points for financial dimension, 85 points for technical dimension, 78 points for market dimension, 80 points for management dimension, and 88 points for innovation dimension (all full marks are 100 points). These quantitative indicators provide the basis for the subsequent classification and ranking of enterprises.

[0052] Step S4: Enterprise classification, scoring and sorting. This step includes three sub-processes: enterprise classification, comprehensive scoring and sorting. First, based on the multi-dimensional quantitative indicators obtained in step S3, machine learning algorithms (such as K-means clustering, decision tree, random forest, etc.) are used to classify the enterprises related to the industrial chain. The results include core enterprises, important suppliers, general suppliers, potential enterprises, etc. Then, each enterprise is comprehensively scored according to the pre-set weights. The scoring formula can be: comprehensive score = w1*financial score + w2*technical score + w3*market score + w4*management score + w5*innovation score, where w1~w5 are the weight coefficients of each dimension, and satisfy w1+w2+w3+w4+w5=1. Finally, the enterprises are sorted in descending order according to the comprehensive score to obtain the final sorting result.

[0053] Step S5: Obtaining the interaction data, transaction intensity data and resource flow data between enterprises. Among them, the interaction data reflects the cooperative relationship between enterprises, such as the number of cooperative projects, strategic alliance relationships, etc.; the transaction intensity data reflects the business transactions between enterprises, such as the purchase amount, sales amount, contract frequency, etc.; the resource flow data reflects the resource exchange between enterprises, such as talent flow, technology transfer, capital transactions, etc. These data can be obtained through questionnaire surveys, public information collection, data purchase, etc.

[0054] Step S6: Construct an intelligent industrial chain network diagram. In this last step, all the data and analysis results obtained in the previous steps are integrated to construct an intelligent industrial chain network diagram. The specific implementation method is as follows: the enterprise is used as a network node, and the node size represents the comprehensive score of the enterprise; the node color represents the enterprise category (for example, core enterprises are red, important suppliers are blue, etc.); the connection between nodes represents the relationship between enterprises, and the thickness of the line represents the transaction intensity or resource flow intensity; the force-directed graph algorithm (Fruchterman-Reingold algorithm) is used to optimize the network layout to make the structure clearer; interactive functions are added to allow users to click on nodes to view detailed information and adjust display parameters. In this way, the final intelligent industrial chain network diagram not only intuitively displays the industrial chain structure, but also contains rich multi-dimensional information, which helps decision makers quickly understand the overall picture of the industrial chain and identify key nodes and potential risks.

[0055] The method in this embodiment can be implemented by a computer program. The program can be stored in a computer-readable storage medium, and when executed by a processor, the above steps are implemented.

[0056] Example 2

[0057] like Figure 2 As shown, in step S1, the process of building a multi-dimensional qualification evaluation model based on industry characteristic data includes the following sub-steps:

[0058] S11: Collect industry characteristic data. First, collect industry characteristic data related to the target industry. These data may include but are not limited to industrial policy information, industry standards, market demand information, and intellectual property information.

[0059] First, industrial policy information, such as industrial development plans and incentive policies issued by the national or local governments, can provide policy support and guidance for the development direction and key areas of the industrial chain. Second, industry standards, including quality standards, technical specifications and safety regulations, help ensure compliance and quality control in all links of the industrial chain. Third, market demand information, such as market size, consumer preferences and demand trends, can help companies understand market dynamics and adjust their production and marketing strategies. In addition, intellectual property information, such as the number of patent applications and the distribution of patent types.

[0060] For example, for the new energy vehicle industry, the industry-specific data that can be collected include the "New Energy Vehicle Industry Development Plan (2021-2035)", "Electric Vehicle Safety Requirements" standards, new energy vehicle market sales data, related patent application statistics, etc. These data provide a rich information basis for subsequent analysis.

[0061] For example, for the artificial intelligence industry, the industry-specific data that can be collected include the "New Generation Artificial Intelligence Development Plan", "Artificial Intelligence Standardization White Paper", global AI chip market size data, machine learning and deep learning related patent application statistics, etc. These data cover policy orientation, technical standards, market demand and innovation trends.

[0062] For example, in the 5G communications industry, the data that can be collected may include the "5G Application "Sailing" Action Plan (2021-2023)", "5G Mobile Communication Network Quality Monitoring Procedures", global 5G base station construction statistics, 5G terminal equipment shipment data, 5G-related core technology patent distribution, etc. These data reflect information such as industry development plans, technical standards, infrastructure construction progress and market acceptance.

[0063] For example, in the new materials industry, the data that can be collected include the "New Materials Industry Development Guide", the "Graphene Material Testing and Characterization" national standard, the global carbon fiber composite material market demand forecast, the application trend report of lightweight materials for new energy vehicles, and the analysis of patents related to high-performance metal materials. These data reflect information such as the industry development direction, technical specifications, market applications, and innovation hotspots.

[0064] For example, for the blockchain technology industry, the industry-specific data that can be collected include the "Blockchain Technology Application Guide", "Blockchain Information Service Management Regulations", the global major countries' digital currency development status report, enterprise-level blockchain solution market size data, distributed ledger technology-related patent application statistics, etc. These data cover policies and regulations, application guidelines, market demand and technological innovation.

[0065] S12: Use natural language processing algorithms to perform text analysis on industry-specific data and extract key features and themes. In this step, natural language processing (NLP) technology is used to analyze the collected text data. It mainly involves technologies such as text preprocessing, topic modeling, keyword extraction, and named entity recognition.

[0066] First, the data is cleaned and normalized through text preprocessing steps, which includes word segmentation, stop word removal, and lemmatization to simplify and unify the text content. Then, topic modeling techniques, such as the Latent Dirichlet Allocation (LDA) model, are used to mine potential topics in the text to reveal the core content and structure of the data. In addition, keyword extraction methods, such as the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm, are used to identify and extract important words in the text to capture the key information of the text. Finally, named entity recognition is performed to identify specific entities in the text, such as company names, technical terms, etc., laying the foundation for further analysis and application.

[0067] In the AI ​​industry, key topics include machine learning, deep neural networks, natural language processing, computer vision, reinforcement learning, edge computing, and AI chips. The corresponding key features include algorithm complexity, model accuracy, computing resource requirements, data processing capabilities, interpretability, real-time performance, and energy efficiency. These topics and features can reveal the core technology development direction of the AI ​​industry and the types of resources required.

[0068] In the biopharmaceutical industry, key topics such as gene editing, immunotherapy, personalized medicine, biomarkers, drug delivery systems, bioinformatics and regenerative medicine reflect the industry's cutting-edge research directions and innovation areas. The corresponding key features include clinical trial success rate, drug half-life, bioavailability, target specificity, side effect level, production cost and patent protection period, which help measure the effectiveness and commercial feasibility of new therapies and technologies.

[0069] For the new energy vehicle industry, key topics include battery technology, charging infrastructure, intelligent networking, autonomous driving, lightweight materials, energy management systems and electric drive systems. Its key features include driving range, charging speed, energy density, safety performance indicators, intelligence level, manufacturing cost and environmental protection index, which reveal the balance point between technological progress and market demand in the industry.

[0070] In the financial technology industry, blockchain technology, digital currency, big data analysis, algorithmic trading, identity authentication, open banking and regulatory technology constitute the key themes. The key features related to them are transaction processing speed, system security level, user authentication accuracy, data encryption strength, risk assessment accuracy, compliance indicators and user experience scores. These features reflect the innovation ability of financial technology and its attention to security and compliance.

[0071] Key topics in the renewable energy industry include solar photovoltaics, wind power generation, hydrogen energy technology, energy storage systems, smart grids, biomass energy and geothermal energy utilization. Its key features include energy conversion efficiency, carbon emission reduction, energy storage capacity, system stability, cost-effectiveness, resource utilization and environmental adaptability.

[0072] In the 5G communications industry, key topics include network slicing, edge computing, millimeter wave technology, massive MIMO, ultra-reliable low-latency communications, network virtualization, and spectrum efficiency. Its key features include transmission rate, network capacity, latency, energy efficiency, coverage, connection density, and network security.

[0073] Key themes in the smart manufacturing industry, such as industrial Internet of Things, digital twins, predictive maintenance, collaborative robots, additive manufacturing (3D printing), virtual reality / augmented reality, and supply chain optimization, reveal the industry's trend towards intelligent and digital transformation. Related key features include production line automation, product defect rate, equipment utilization, production efficiency, customization capabilities, energy consumption, and data integration.

[0074] As can be seen above, each industry has its own unique key themes and quantifiable characteristics. This information can be extracted through NLP analysis of text data such as industry reports, policy documents, and research papers. This analysis can not only help understand the core development direction and technical focus of each industry, but also provide specific and measurable indicators for building an accurate qualification assessment model.

[0075] S13: Based on the extracted key features and themes, combined with the expert knowledge base, determine the target dimensions of qualification evaluation, and set specific evaluation indicators and weights for each target dimension. In this step, the NLP analysis results are combined with the knowledge and experience of industry experts to determine the target dimensions of qualification evaluation. These dimensions should fully reflect the comprehensive strength and competitiveness of the enterprise in the industry.

[0076] In the biopharmaceutical industry, the target dimensions of qualification assessment cover R&D investment, product pipeline, clinical trial capabilities, manufacturing capabilities, regulatory compliance, intellectual property portfolio, and industry-university-research cooperation. Through these dimensions, the company's comprehensive capabilities in R&D funding, the number of products under development and market potential, clinical trial management efficiency, production quality control, regulatory compliance, patent protection, and the depth and breadth of cooperation with research institutions can be examined.

[0077] For the new energy vehicle industry, the target dimensions of qualification assessment include battery technology level, intelligence level, manufacturing capacity, supply chain management, charging facility layout, brand influence and environmental performance. These dimensions help evaluate the company's strength in battery research and development, autonomous driving and intelligent network technology applications, production capacity, supply chain stability, investment and layout of charging infrastructure, market awareness, and carbon emissions and recyclability during the product life cycle.

[0078] In the financial technology industry, the target dimensions of qualification assessment include technological innovation capabilities, risk control capabilities, data security, compliance level, user experience, market penetration and cross-border integration capabilities. Through these dimensions, the company's performance and capabilities in cutting-edge technology applications, risk management, data protection, financial regulatory compliance, customer experience, market share and cooperation with traditional financial institutions can be measured.

[0079] The target dimensions for qualification assessment in the renewable energy industry include technical efficiency, cost control capability, project development capability, grid friendliness, environmental impact, policy responsiveness and energy storage integration. These dimensions assess the company's comprehensive performance in terms of energy conversion efficiency, cost competitiveness, new project development, grid adaptability, environmental impact, policy responsiveness and energy storage technology application.

[0080] For the 5G communications industry, the target dimensions of qualification assessment include technological innovation capabilities, network deployment efficiency, spectrum resources, equipment R&D capabilities, network security, industry application expansion and internationalization. These dimensions are used to assess the company's capabilities and achievements in 5G core technology innovation, network construction, spectrum resources, independent equipment R&D, network security protection, industry application promotion and global market expansion.

[0081] In the intelligent manufacturing industry, the target dimensions of qualification assessment include digitalization level, flexible manufacturing capabilities, degree of intelligence, quality control, energy efficiency, supply chain integration and talent structure. These dimensions assess the company's comprehensive strength in terms of digitalization of production processes, flexible manufacturing capabilities that respond to market changes, application of intelligent technologies, product quality control, energy efficiency, collaboration with upstream and downstream supply chains, and talent team building.

[0082] These dimensions provide a comprehensive qualification assessment framework for each industry, covering multiple aspects such as technology, management, market and compliance. In practical applications, the most relevant dimensions can be selected for key assessment based on the characteristics of specific enterprises and the purpose of assessment, and appropriate weights can be assigned to each dimension. At the same time, these dimensions should also be continuously updated and adjusted with the development of the industry and the advancement of technology.

[0083] For each identified target dimension, set a number of specific evaluation indicators and assign appropriate weights to each indicator. The weight allocation should take into account the degree of influence of each indicator on the overall strength of the enterprise. For example, for the dimension of "technological innovation capability", the evaluation indicators and weights include the proportion of R&D investment (30%), the number of invention patents (25%), the speed of new product launch (20%), the proportion of technical personnel (15%) and the number of industry-university-research cooperation projects (10%). This detailed indicator system can more accurately measure the performance of enterprises in various aspects.

[0084] S14: Use machine learning algorithms to integrate the evaluation indicators and weights into a multi-dimensional qualification evaluation model. In this step, a machine learning algorithm (such as random forest, support vector machine or neural network) is used to integrate the previously determined evaluation indicators and weights into a unified multi-dimensional qualification evaluation model. This model can receive the multi-dimensional data of the enterprise as input and output the comprehensive qualification score of the enterprise.

[0085] The training process of the multi-dimensional qualification assessment model includes the following key steps. The first is data preparation, which collects the data of various indicators of known enterprises and their corresponding comprehensive evaluation results to provide sufficient training samples for the model. These data lay the foundation for the training and evaluation of the model. The next step is data partitioning, which randomly divides the collected data set into training set, validation set and test set according to a certain ratio (such as 6:2:2), so as to ensure that the model has enough sample size to effectively evaluate its performance during the training, validation and testing stages. Then, feature engineering is performed to preprocess and transform the training set data to ensure that the data format and features are suitable for the input requirements of the model. At the same time, the same preprocessing and transformation methods are applied to the validation set and test set to ensure data consistency and the accuracy of model evaluation. On this basis, model selection is performed, and the performance of different machine learning algorithms and model structures is compared by using the processed training data, and the model that best solves the specific problem is selected. After that, the model training stage is entered, and the selected model is trained using the processed training data so that it can learn the laws and patterns in the data and improve the model's predictive ability. Next comes model validation and optimization, where the model performance is evaluated using the validation set, and necessary parameter adjustments and optimizations are performed based on the evaluation results to ensure that the model has good generalization ability and accuracy. Finally, the final test phase is the final optimized model, which is fully evaluated using the test set. This step is designed to test the effectiveness and reliability of the model in actual applications, ensure its stability and accuracy in different scenarios, and provide confidence for the final deployment of the model.

[0086] Through this process, this embodiment constructs a multi-dimensional model that can comprehensively evaluate the qualifications of enterprises. It not only takes into account the particularity of the industry, but also integrates expert knowledge and data-driven machine learning methods to provide a reliable enterprise qualification evaluation basis for the subsequent industrial chain construction.

[0087] Example 3

[0088] like Figure 3 As shown, in a specific embodiment of the present invention, step S2 includes the following sub-steps:

[0089] First, step S21 is to obtain multi-dimensional qualification data of enterprises related to the industrial chain. This step aims to comprehensively collect various data reflecting the qualifications of enterprises, including internal enterprise data, public data, third-party data, and Internet of Things data. The internal data of enterprises is obtained by collecting financial data and human resource data from the enterprise resource planning (ERP) system, financial management system, and human resource management system to reflect the financial health, talent structure, and management efficiency of the enterprise. Specific data include operating income, profit margin, debt-to-asset ratio, cash flow, total number of employees, educational distribution, proportion of technical personnel, and staff turnover rate. Public data is obtained from government public databases, industry association websites, and corporate annual reports to obtain market share data and environmental compliance data to reflect the market position and social responsibility fulfillment of enterprises. Specific data include product market share, sales area distribution, customer satisfaction, environmental protection investment, carbon emissions, waste treatment, environmental certification, etc. Third-party data acquisition obtains technology innovation data and supply chain management data through a third-party data service platform to reflect the innovation ability and supply chain efficiency of enterprises. Specific data include the number of patent applications, R&D investment ratio, new product development cycle, supplier rating, raw material procurement cost, inventory turnover rate, etc. IoT data is acquired from the IoT sensor network deployed on the company's production line, which collects production capacity data from related companies in the industrial chain to reflect the company's production efficiency and capacity utilization in real time. Specific data include equipment operating efficiency, product yield, production line capacity utilization, etc.

[0090] Step S22 is data processing, which involves cleaning, verifying and standardizing the acquired multi-dimensional qualification data to generate a structured multi-dimensional qualification data set with a unified format and specifications. Data processing includes four specific steps:

[0091] The first step is data cleaning, which ensures the purity and integrity of the data by removing duplicate data, processing missing values, and detecting outliers. For example, use data deduplication algorithms to identify and delete duplicate records, use mean filling, median filling, or advanced interpolation methods to handle missing data based on data characteristics, and use statistical methods (such as the Z-score method) or machine learning algorithms (such as isolation forests) to identify and process outliers.

[0092] The second is data validation, which ensures the accuracy and consistency of data through cross-validation, logic validation and time series validation. Specific methods include cross-comparing data from different sources, checking whether the data conforms to the predetermined business logic and constraints, and checking the continuity of time series data and the rationality of the trend.

[0093] Next, the data is standardized to ensure consistency and comparability in subsequent analysis. The steps of standardization include unit unification, converting data of different units to a unified standard, such as unifying financial data into RMB; format normalization, unifying date format, numerical precision, etc. to ensure consistency of data format; data normalization, using methods such as Min-Max normalization or Z-score normalization to convert data of different dimensions to the same scale.

[0094] Finally, structured processing is performed to design a database model suitable for multi-dimensional qualification data, including dimension tables and fact tables. The processed data is then loaded into the database according to the designed model, and index optimization is performed to improve the efficiency of subsequent data query and analysis.

[0095] Through the above steps, a structured multi-dimensional qualification data set with unified format and standardization is obtained, which provides a reliable data basis for the subsequent qualification evaluation model and ensures the accuracy and credibility of the evaluation results. In practical applications, the source and processing methods of data acquisition can be appropriately adjusted according to the specific industry characteristics and data availability.

[0096] Example 4

[0097] like Figure 4 As shown, in one embodiment, step S3 includes: S31: inputting the multidimensional qualification data into the multidimensional qualification evaluation model; S32: using the multidimensional qualification evaluation model to process and calculate the input multidimensional qualification data, specifically including: for each target dimension, applying the corresponding evaluation indicators and weights, and calculating the score of a single target dimension: Si = Σ(wij*xij); wherein Si is the score of the i-th target dimension, wij is the weight of the j-th indicator under the i-th target dimension, and xij is the standardized value of the j-th indicator under the i-th dimension; calculating the overall qualification evaluation score: S = Σ(Wi*Si), wherein S is the overall qualification evaluation score, and Wi is the weight of the i-th target dimension; outputting the multidimensional quantitative indicators of the qualification data of related enterprises in the industrial chain, wherein the multidimensional quantitative indicators include the specific indicator values ​​of each target dimension and the overall qualification evaluation score.

[0098] In this embodiment, step S31 involves inputting multi-dimensional qualification data into a multi-dimensional qualification evaluation model. The multi-dimensional qualification data here may include data on the company's financial status, technological innovation capabilities, market share, human resource quality, and other aspects. For example, for a manufacturing company, its multi-dimensional qualification data may include indicators such as annual turnover, R&D investment ratio, market share of major products, and proportion of highly educated employees. These data are input into the multi-dimensional qualification evaluation model through a pre-set data interface or data input module.

[0099] Step S32 is the core step of processing and calculating the input multi-dimensional qualification data using the multi-dimensional qualification evaluation model. In this step, the corresponding evaluation indicators and weights are first applied to each target dimension to calculate the score of a single target dimension. The target dimensions here can be the financial status, technological innovation capability, etc. mentioned above. Taking technological innovation capability as an example, its evaluation indicators may include the proportion of R&D investment, the number of patent applications, the new product development cycle, etc., and each indicator has a corresponding weight. The score of this dimension is calculated by the formula Si=Σ(wij*xij), where wij represents the weight of each indicator, and xij represents the standardized value of each indicator. The standardization process is to eliminate the dimensional differences between different indicators so that the indicators are comparable.

[0100] After calculating the scores of each target dimension, the next step is to calculate the overall qualification assessment score. This is achieved through the formula S = Σ(Wi*Si), where Wi represents the weight of each target dimension. These weights are set based on industry characteristics and assessment purposes. For example, for high-tech companies, the weight of technological innovation capabilities may be higher. Through this weighted summation method, a quantitative indicator that comprehensively reflects the overall qualification level of the enterprise can be obtained.

[0101] Finally, the model outputs multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain. These indicators include two levels: one is the specific indicator value of each target dimension, and the other is the overall qualification evaluation score. Such output results can not only allow users to understand the performance of enterprises in various specific aspects, but also quickly compare the comprehensive strength of different enterprises through the overall score.

[0102] Example 5

[0103] like Figure 5 As shown, in a specific embodiment of the present invention, step S4 includes the following sub-steps:

[0104] In step S41, the K-means clustering algorithm is used to perform cluster analysis on related enterprises in the industrial chain. First, the multi-dimensional quantitative indicators are standardized using the Z-score method: z = (x-μ) / σ, where x is the original value, μ is the mean, and σ is the standard deviation. This ensures that indicators of different scales can be compared in the same space. Then, the K-means algorithm is implemented using the KMeans class in the sklearn library. The selection of the number of categories K adopts the Elbow Method: calculate the within-cluster sum of squares (WSS) under different K values ​​(such as 2 to 20), draw a curve of WSS changing with K, and select the K value corresponding to the inflection point of the curve. In order to improve the stability of clustering, the K-means++ method is used to initialize the cluster center, and n_init = 10, max_iter = 300 are set to ensure that the algorithm fully converges. After clustering is completed, the number of enterprises in each category is calculated, and the category center point is used as the category feature. The relative relationship between categories is expressed by calculating the Euclidean distance between the category center points. Finally, the enterprise category distribution information is stored in a dictionary, which contains the enterprise ID list of each category, the category feature vector and the distance matrix between categories.

[0105] In step S42, the comprehensive qualification index of the enterprise is calculated. First, a weight is set for each target dimension, which can be determined by expert evaluation or importance analysis based on historical data. Assuming there are n target dimensions, the weight vector is W = [w1, w2, ..., wn]. For each enterprise, its score on each dimension forms a score vector S = [s1, s2, ..., sn]. The calculation of the comprehensive qualification index adopts the weighted average method: CI = Σ(wi*si) / Σwi, i from 1 to n. In order to deal with possible missing values, the following strategy is adopted: if the score of a dimension is missing, the dimension and its weight are ignored in the calculation. In order to ensure the comparability of the comprehensive indicators of different enterprises, the calculation results are normalized using the Min-Max scaling method: CI_norm = (CI-CI_min) / (CI_max-CI_min), where CI_min and CI_max are the minimum and maximum values ​​of the comprehensive indicators of all enterprises, respectively. In this way, the final comprehensive qualification index is scaled to the interval [0,1].

[0106] In step S43, the enterprises in each category are sorted according to the comprehensive index. First, the enterprises are grouped according to the categories obtained in step S41. For each category, Python's sorted() function is used to sort all enterprises in the category in descending order with the comprehensive qualification index CI_norm as the key. In order to provide more detailed ranking information, the percentile of each enterprise in its category is calculated: Percentile = (Rank-1) / (N-1)*100, where Rank is the ranking of the enterprise in the category (starting from 1), and N is the total number of enterprises in the category. This percentile reflects the relative position of the enterprise in the same category. At the same time, in order to facilitate subsequent analysis and display, this embodiment also calculates the quartiles (Q1, Q2, Q3) in each category and the standard deviation of the enterprise score. Finally, the sorting results are stored in the form of a dictionary, with the key being the category ID and the value being a list containing the enterprise ID, the comprehensive index score, the intra-class ranking, and the percentile. This structured sorting result facilitates subsequent data analysis and visualization, such as generating a score distribution map within a category or identifying the top-N enterprises in each category.

[0107] Example 6

[0108] In a specific embodiment of the present invention, step S5 includes the following process in detail:

[0109] The first step is to obtain supply relationship data, which is defined as information reflecting the transactions and degree of dependence between an enterprise and its suppliers, including purchase volume, purchase frequency, and supplier importance. The acquisition process involves connecting to the enterprise resource planning (ERP) system database, extracting basic supplier information (such as supplier ID, name, contact information), purchase order records (such as purchase category, quantity, amount, date), and supplier evaluation data (such as supplier grade, delivery timeliness, quality qualification rate); at the same time, accessing the supply chain management system to obtain supplier inventory levels, supplier delivery cycles, and supply chain risk assessment data; and then integrating data from ERP and supply chain management systems to form a complete supply relationship data set.

[0110] The second step is to obtain partnership data, which reflects information about the cooperation between enterprises and other enterprises in terms of projects, R&D, and markets, including the frequency, intensity, and effectiveness of cooperation. The acquisition process includes accessing the enterprise cooperation project database, extracting basic project information (such as project name, type, participants, start and end time), project scale (such as investment amount, personnel input), and project progress and results data; at the same time, connecting to the contract management system to obtain basic contract information (such as contract type, contracting party, contract amount, validity period), contract fulfillment status, and payment records; and finally integrating project data and contract data to build a complete partnership data set.

[0111] The third step is to obtain competitive relationship data, which reflects the relative position and competitive situation between the company and its competitors in the market, including market share, product comparison, customer evaluation and other information. The acquisition process involves collecting and analyzing market research reports, extracting market share data, brand awareness and customer satisfaction survey results; at the same time, studying industry analysis reports, obtaining industry competition landscape descriptions, major competitor information and market concentration data; and then integrating competition data from different reports to form a competitive relationship data set.

[0112] Finally, we need to obtain technology dependency data, which is defined as information reflecting technology innovation, patent citations, and technology licensing between enterprises, including the number of patents, citation relationships, and technology transfers. The acquisition process includes querying the patent database, extracting patent application and authorization information, patent categories, and patent citation relationships; analyzing technology licensing agreements, obtaining licensor and licensee information, technology licensing scope, and licensing fees; and finally integrating patent data and technology licensing information to construct a technology dependency dataset.

[0113] In actual operation, the above data can be automatically obtained by writing data crawling scripts or using API interfaces. For data that cannot be directly obtained, manual collection and sorting can be adopted. After obtaining the data, data cleaning, standardization and structured processing are required to ensure data quality and consistency. Through the above process, the supply, cooperation, competition and technology dependency data between related enterprises in the industrial chain can be comprehensively collected. These multi-dimensional relationship data provide the necessary information basis for the subsequent construction of a complete industrial chain network, which helps to deeply analyze the complex interactions and dependencies between enterprises.

[0114] Example 7

[0115] like Figure 6 As shown, in a preferred embodiment of the present invention, step S6 includes the following process in detail:

[0116] In step S61, the establishment of the intelligent matching algorithm adopts the random forest model. First, multi-dimensional quantitative indicators (such as financial indicators, technological innovation indicators, market position indicators, etc.), enterprise category distribution (category labels obtained by K-means clustering), sorting results (ranking within the category) and industry characteristic data (such as industry growth rate, market concentration, etc.) are used as feature inputs. Then, historical successful cooperation cases are used as training sets, which contain about 10,000 pairs of successful cooperation companies. In the feature engineering stage, categorical variables are uniquely encoded and numerical variables are standardized. The random forest model settings include: the number of trees is 500, the maximum depth is 10, and the minimum number of leaf node samples is 5. The hyperparameters are optimized using 5-fold cross validation, and the final model achieves an accuracy of 85% on the test set.

[0117] In step S62, a genetic algorithm is used for optimal pairing. The chromosome encoding uses integer encoding, each gene represents an enterprise ID, and the chromosome length is equal to the total number of enterprises. The fitness function is defined as: f = Σ(wij*pij), where wij is the collaboration weight between enterprises i and j (predicted by the random forest model), and pij is the pairing indicator variable. The population size is set to 1000, the crossover probability is 0.8, and the mutation probability is 0.1. The parent generation is selected using the tournament selection method, using single-point crossover and random mutation operations. Stop when there is no improvement after 500 iterations or 50 consecutive generations. Finally, the chromosome with the highest fitness is output as the optimal pairing solution.

[0118] When constructing the first industrial chain network diagram in step S63, the NetworkX library is used. Each enterprise is created as a node, and the node attributes include: {'id': enterprise ID, 'category': cluster category, 'rank': ranking within the category, 'metrics': {multi-dimensional quantitative indicator dictionary}}. The initial collaboration relationship is added to the graph as an edge, and the edge attributes include: {'weight': initial collaboration weight}. The ForceAtlas2 algorithm is used for layout optimization to ensure visual clarity.

[0119] In step S64, when integrating the interaction data, first construct an inter-enterprise interaction matrix I, where Iij represents the interaction intensity between enterprises i and j. The interaction intensity is calculated by the following formula: Iij = α*transaction frequency + β*log(transaction amount) + γ*number of cooperative projects, where α, β, and γ are weight coefficients determined by expert evaluation. Then update the edge attributes in the network graph: {'type': determined according to the main interaction type, 'weight': original weight*(1+log(1+Iij))}.

[0120] Finally, in step S65, the calculation formula for collaboration strength CS is: CSij = (Tij / max(Ti))*(1+log(1+Vij)), where Tij is the annual transaction amount between enterprises i and j, Ti is the total transaction amount of enterprise i, and Vij is the transaction frequency. The calculation of resource flow RF considers multiple dimensions: RFij = w1*capital flow+w2*talent flow+w3*technology transfer, where w1, w2, and w3 are weight coefficients. Add these newly calculated attributes to the attributes of the edge: {'collaboration_strength':CSij,'resource_flow':RFij}. The final network diagram is visualized using the PyViz library. The size of the node reflects the size of the enterprise, the color represents the category, the thickness of the edge represents the collaboration strength, and the color depth of the edge represents the resource flow.

[0121] Through the above solution, a comprehensive and dynamic intelligent industrial chain network diagram is constructed, which can accurately reflect the complex relationships and resource flows between enterprises.

[0122] Furthermore, in step S611, when constructing the feature vector space, each enterprise is represented as an n-dimensional feature vector. Specifically, assuming that the multi-dimensional quantitative indicators include m indicators (such as operating income growth rate, R&D investment ratio, market share, etc.), the first m dimensions of the vector correspond to these indicator values. The category to which the enterprise belongs is encoded using unique hot encoding. Assuming there are k categories, k dimensions are used to represent the category information. The last dimension is used to represent the ranking percentage of the enterprise in its category. Therefore, the dimension n of the feature vector is n=m+k+1. In order to ensure that the values ​​of different dimensions are comparable, all numerical features are standardized using the Z-score standardization method: z=(x-μ) / σ, where x is the original value, μ is the mean of the feature, and σ is the standard deviation.

[0123] Furthermore, in step S612, the similarity calculation method is determined and implemented as follows: First, the Euclidean distance between the multi-dimensional feature vectors is calculated as the first similarity value. Assuming that the feature vectors of the two enterprises are a and b, the Euclidean distance i ranges from 1 to n. The first similarity value S1=1 / (1+d), ensuring that the similarity is between 0 and 1. Next, the predefined enterprise category compatibility matrix C is used, where Cij represents the compatibility weight (between 0 and 1) of category i and category j. Assuming that enterprise A belongs to category i and enterprise B belongs to category j, the compatibility weight w=Cij. The second similarity value S2=S1*w. Then, calculate the ranking influence factor f=1-|PA-PB| / 2, where PA and PB are the ranking percentages of the two enterprises in their respective categories, respectively. The third similarity value S3=S2+(1-S2)*f. Finally, the final similarity is calculated by weighted average: S=α*S1+β*S2+γ*S3, where α, β, γ are weight coefficients, and α+β+γ=1. These weights can be optimized by cross-validation.

[0124] Among them, the enterprise category compatibility matrix C is a predefined n×n square matrix used to quantify the compatibility or correlation between different enterprise categories. Each element Cij in the matrix represents the compatibility weight between category i and category j, ranging from 0 to 1, where 0 indicates complete incompatibility and 1 indicates complete compatibility. Usually, the matrix C is symmetric (Cij=Cji), reflecting the bidirectional nature of compatibility between categories. The diagonal element Cii of the matrix is ​​1, indicating complete compatibility between similar enterprises. This matrix can be constructed in a variety of ways, including expert evaluation (industry experts assign compatibility scores based on experience), data-driven methods (calculating cooperation frequency or success rate based on historical cooperation data or transaction data), and text analysis (analyzing industry reports, news and other texts to calculate the co-occurrence frequency of different enterprise categories). By using the enterprise category compatibility matrix, the inherent correlation between enterprises of different categories can be considered in the enterprise matching process, thereby improving the accuracy and practicality of industrial chain network construction.

[0125] Further, in step S613, a gradient boosting decision tree (GBDT) is selected as a machine learning-based enterprise matching model. The XGBoost library is used for specific implementation. The input features include: the concatenation of the multidimensional feature vectors of the two enterprises (2n dimensions) and the similarity results between them (1 dimension). The label is binary, 1 indicates a successful match, and 0 indicates a failed match. The training data set contains successful and failed enterprise matching cases in history, with a ratio of 1:3. The model parameters are set as follows: the maximum tree depth is 6, the learning rate is 0.1, the subsampling rate is 0.8, the feature sampling rate is 0.8, the minimum number of child node samples is 5, and the number of iterations is 1000. Using the early stopping strategy, the training is stopped when the AUC on the validation set does not improve within 50 iterations. In order to deal with the problem of class imbalance, the class weight adjustment is adopted, and the weight of the positive sample is set to 3 and the weight of the negative sample is 1. Finally, the model outputs the matching probability between the two enterprises, ranging from 0 to 1. By setting a threshold (for example, 0.5), the probability can be converted into a binary decision of whether to match or not. The model was evaluated using 5-fold cross validation, and key indicators included AUC (area under the curve), precision, recall, and F1 score.

[0126] This model not only takes into account the multidimensional characteristics of enterprises, the compatibility and ranking information of enterprise categories, but also uses machine learning algorithms to capture complex nonlinear relationships, so that it can effectively predict the matching probability between enterprises and provide reliable decision-making support for industrial chain collaboration.

[0127] Example 8

[0128] like Figure 7As shown, in a preferred embodiment of the present invention, it also includes S7: based on the intelligent industrial chain network diagram, intelligent diagnosis of the intelligent industrial chain network is performed to identify potential problems. Wherein, step S7 includes the following processes in detail: S71: using the social network analysis algorithm to analyze the intelligent industrial chain network diagram, and calculating the following network indicators: node centrality, edge weight, clustering coefficient, network density; S72: based on the network indicators, identifying key node enterprises and potential risk connections, which includes: identifying the first type of node enterprises with node centrality higher than the preset threshold as key node enterprises; identifying the connection with edge weight lower than the preset threshold as a potential risk connection; identifying the second type of node enterprises whose clustering coefficient deviates from the average clustering coefficient of the industrial chain network, wherein the average clustering coefficient of the industrial chain network is obtained by calculating the arithmetic mean of the clustering coefficients of all nodes in the intelligent industrial chain network diagram; S73: based on the network indicators, the key node enterprises and the potential risk connection, analyzing and obtaining the potential chain breaking risk; S74: generating an intelligent diagnosis report, including: detailed information of the first type of node enterprises, enterprises related to the potential risk connection and the second type of node enterprises; and the potential chain breaking risk.

[0129] In step S71, the social network analysis algorithm is implemented using the NetworkX library to calculate network indicators. Node centrality uses three indicators: degree centrality, betweenness centrality, and eigenvector centrality. Degree centrality DC(v) = deg(v) / (N-1), where deg(v) is the degree of node v and N is the total number of nodes. Betweenness centrality where σ st is the number of shortest paths from s to t, σ st (υ) is the number of shortest paths from s to t that pass through v. Eigenvector centrality is calculated by solving the equation Ax = λx, where A is the adjacency matrix, x is the eigenvector, and λ is the largest eigenvalue. Edge weights are extracted directly from the network graph. Clustering coefficient C(v) = 2e / (k(k-1)), where e is the number of edges between neighbors of node v and k is the degree of v. Network density D = 2m / (n(n-1)), where m is the number of edges and n is the number of nodes. These calculations are stored in a dictionary with the key being the node ID and the value being a dictionary containing the metrics.

[0130] In step S72, key node enterprises are first identified. The thresholds θ1, θ2, and θ3 are set to the 95th percentiles of degree centrality, betweenness centrality, and eigenvector centrality, respectively. If any centrality index of a node exceeds the corresponding threshold, it is marked as a first-class key node enterprise. For potential risk connections, the edge weight threshold θ4 is set to the 10th percentile of all edge weights, and edges below this value are marked as potential risk connections. Calculate the average clustering coefficient C_avg = ΣC(v) / n of the industrial chain network. Define the clustering coefficient deviation δ = |C(v)-C_avg| / C_avg, and set the threshold θ5 = 0.5. If δ>θ5 of node v, it is marked as a second-class node enterprise. All identification results are stored in a structured JSON object, which contains a list of various types of enterprises and their related indicators.

[0131] In step S73, the potential risk of chain break is analyzed based on network indicators, key node enterprises and potential risk connections. First, a risk scoring model is constructed: R = w1*(1-D)+w2*(1-C_avg)+w3*(N1 / N)+w4*(E_risk / E), where D is the network density, C_avg is the average clustering coefficient, N1 is the number of first-class node enterprises, N is the total number of nodes, E_risk is the number of potential risk connections, E is the total number of edges, and w1 to w4 are weight coefficients. Then, the connectivity of the network is analyzed using a graph theory algorithm. The articulation points and bridges of the network are calculated, which are nodes and edges that will cause the network to split after removal. For each articulation point and bridge, the increase in the number of network components after its removal is calculated as the degree of its chain break impact. Finally, a ranked list is generated based on the risk score and the degree of chain break impact, identifying the key nodes and connections that are most likely to cause chain breakage.

[0132] In step S74, an intelligent diagnosis report is generated. The report is in HTML format and includes the following main parts: Network overview, which displays basic information such as network density, average clustering coefficient, total number of nodes and total number of edges. Key node enterprise analysis, which lists the first type of node enterprises, including their ID, name, centrality index and description of their importance in the network. Potential risk connection analysis, which lists all potential risk connections, including the enterprise information at both ends of the connection, edge weights and possible risk causes. Abnormal aggregation enterprise analysis, which lists the second type of node enterprises, including their clustering coefficient, deviation and possible cause analysis. Potential chain break risk assessment, which displays the overall risk score, lists the top-N (e.g., top-10) nodes and connections that are most likely to cause chain breaks, and provides corresponding risk mitigation recommendations. Visualization part, which uses the Plotly library to generate an interactive network diagram, highlighting key nodes, risk connections and abnormal aggregation areas. The report is generated by the Jinja2 template engine to ensure dynamic updating of content and consistency of format.

[0133] Through this series of in-depth network analysis and intelligent diagnosis, we can comprehensively assess the health of the industrial chain network, identify potential risk points and vulnerable links, and provide valuable advice to decision makers, thereby improving the resilience and sustainability of the industrial chain.

[0134] Example 9

[0135] like Figure 8 As shown, a smart industrial chain construction system 200 based on multi-dimensional qualification certification includes:

[0136] The qualification evaluation model construction module 210 is used to construct a multi-dimensional qualification evaluation model based on industry characteristic data;

[0137] The data acquisition module 220 is used to obtain the pre-processed multi-dimensional qualification data of the enterprises related to the industrial chain;

[0138] The qualification analysis module 230 is used to analyze the multi-dimensional qualification data using the multi-dimensional qualification evaluation model to obtain multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain;

[0139] An enterprise classification and ranking module 240 is used to classify the enterprises related to the industrial chain based on the multi-dimensional quantitative indicators of the qualification data using a machine learning algorithm to obtain enterprise category distribution, and to comprehensively score and rank the enterprises related to the industrial chain according to the set weights to obtain ranking results;

[0140] Interaction data acquisition module 250, used to acquire interaction data, transaction intensity data and resource flow data between related enterprises in the industrial chain;

[0141] The network diagram construction module 260 is used to construct an intelligent industrial chain network diagram based on the multi-dimensional quantitative indicators, the enterprise category distribution, the sorting results, the industry characteristic data, the interaction data, the transaction intensity data and the resource flow data.

[0142] In some embodiments, the qualification assessment model building module 210 includes:

[0143] An industry characteristic data acquisition unit, used to acquire industry characteristic data including industrial policy information, industry standards, market demand information and intellectual property information;

[0144] A text analysis unit, used to perform text analysis on the industry characteristic data using a natural language processing algorithm to extract key features and themes related to the industry;

[0145] A target dimension determination unit, used to determine the target dimension of qualification evaluation based on the expert knowledge base, the key features and the subject, and set evaluation indicators and weights for each target dimension, wherein the weights reflect the relative importance of each dimension in industrial development;

[0146] The model fusion unit is used to adopt a machine learning algorithm to normalize and weightedly combine the evaluation indicators and weights of different target dimensions, and fuse them into a unified multi-dimensional qualification evaluation model.

[0147] In some embodiments, the qualification analysis module 230 includes:

[0148] A data input unit, used to input the multi-dimensional qualification data into the multi-dimensional qualification assessment model;

[0149] The data processing and calculation unit is used to process and calculate the input multi-dimensional qualification data using the multi-dimensional qualification evaluation model, specifically including:

[0150] For each target dimension, apply the corresponding evaluation indicators and weights to calculate the score of a single target dimension: Si = Σ(wij*xij); where Si is the score of the i-th target dimension, wij is the weight of the j-th indicator under the i-th target dimension, and xij is the standardized value of the j-th indicator under the i-th dimension;

[0151] Calculate the overall qualification assessment score: S = Σ(Wi*Si), where S is the overall qualification assessment score and Wi is the weight of the i-th target dimension;

[0152] The output unit is used to output multi-dimensional quantitative indicators of the qualification data of enterprises related to the industrial chain, and the multi-dimensional quantitative indicators include specific indicator values ​​of each target dimension and an overall qualification assessment score.

[0153] In some embodiments, the enterprise classification sorting module 240 includes:

[0154] A cluster analysis unit, which is used to perform cluster analysis on related enterprises in the industrial chain based on multi-dimensional quantitative indicators using a K-means clustering algorithm to obtain enterprise category distribution, wherein the enterprise category distribution includes the number of enterprises in each category, category characteristics, and relative relationships between categories;

[0155] A comprehensive index calculation unit is used to perform weighted average of multi-dimensional quantitative indicators according to the weight set for each target dimension to obtain the comprehensive index of the enterprise. The calculation formula of the comprehensive qualification index is as follows: comprehensive qualification index = Σ(Wi*Si) / Σ(Wi), wherein Wi is the weight of the i-th target dimension, and Si is the score of the i-th target dimension;

[0156] The sorting unit is used to sort the enterprises in each category according to the comprehensive index to obtain the sorting result.

[0157] In some embodiments, the network diagram construction module 260 includes:

[0158] An intelligent matching algorithm establishing unit, used to establish an intelligent matching algorithm based on the multi-dimensional quantitative indicators, the enterprise category distribution, the ranking results and the industry characteristic data;

[0159] An enterprise matching unit, used to use the intelligent matching algorithm to optimally match related enterprises in the industrial chain to obtain a preliminary industrial chain collaboration relationship network;

[0160] A preliminary network diagram construction unit is used to construct a first industrial chain network diagram based on the preliminary industrial chain collaborative relationship network, wherein enterprise nodes represent various enterprises in the industrial chain, relationship edges represent collaborative relationships between enterprises, and attributes of enterprise nodes include multi-dimensional quantitative indicators, enterprise category distribution and ranking results;

[0161] A network graph updating unit, used to integrate the acquired interaction data between related enterprises in the industrial chain into the first industrial chain network graph, update the relationship edge type and relationship strength weight in the first industrial chain network graph, and obtain a second industrial chain network graph;

[0162] The network diagram improvement unit is used to calculate the collaboration intensity between enterprises based on the transaction intensity data, calculate the resource flow between enterprises based on the resource flow data, and add the collaboration intensity and the resource flow to the attribute information of the second industrial chain network diagram to form a final intelligent industrial chain network diagram.

[0163] In some embodiments, the intelligent matching algorithm establishing unit includes:

[0164] A feature vector space construction subunit is used to construct a feature vector space, and represent each enterprise as a multi-dimensional feature vector, wherein the multi-dimensional feature vector includes: the value of each indicator in the multi-dimensional quantitative indicator, the code of the category to which the enterprise belongs, and the ranking percentage of the enterprise in the category to which it belongs;

[0165] A similarity calculation subunit, used to determine a similarity calculation method based on the constructed feature vector space, and calculate similarity results between enterprises based on the similarity calculation method;

[0166] The matching model training subunit is used to use the multi-dimensional feature vector and the similarity results between the enterprises to train an enterprise matching model based on machine learning to predict the matching probability between enterprises.

[0167] In some embodiments, the similarity calculation subunit is specifically used for:

[0168] Calculate the Euclidean distance between the multi-dimensional feature vectors to obtain a first similarity value;

[0169] Using a predefined enterprise category compatibility matrix, based on the category code of the enterprise, determining compatibility weights between enterprises of different categories, and multiplying the compatibility weights by the first similarity value to obtain a second similarity value;

[0170] Calculate a ranking influence factor according to the ranking percentage of the enterprise in the category to which it belongs, and use the ranking influence factor to multiply or add the second similarity value to obtain a third similarity value;

[0171] The similarity result between the enterprises is obtained by weighted average calculation based on the first similarity value, the second similarity value and the third similarity value.

[0172] In some embodiments, the matching model training subunit is specifically used for:

[0173] Choose a machine learning-based enterprise matching model from random forest or gradient boosted decision tree;

[0174] The multidimensional feature vector and the similarity results between the enterprises are used as inputs of an enterprise matching model based on machine learning, and the enterprise matching model based on machine learning is trained so that it can predict the matching probability between enterprises.

[0175] In some embodiments, the system further comprises:

[0176] The intelligent diagnosis module is used to perform intelligent diagnosis on the intelligent industrial chain network based on the intelligent industrial chain network diagram and generate an intelligent diagnosis report.

[0177] In some embodiments, the intelligent diagnostic module includes:

[0178] A network analysis unit, used to analyze the intelligent industrial chain network diagram using a social network analysis algorithm, and calculate the following network indicators: node centrality, edge weight, clustering coefficient, and network density;

[0179] A key node identification unit is used to identify key node enterprises and potential risk connections based on the network indicators:

[0180] Identify the first type of node enterprises with node centrality higher than the preset threshold as key node enterprises;

[0181] Identify connections with edge weights below a preset threshold as potential risk connections;

[0182] Identify the second type of node enterprises whose clustering coefficients deviate from the average clustering coefficient of the industrial chain network, wherein the average clustering coefficient of the industrial chain network is obtained by calculating the arithmetic mean of the clustering coefficients of all nodes in the intelligent industrial chain network diagram;

[0183] A risk analysis unit, configured to analyze and obtain a potential link-breaking risk based on the network indicators, the key node enterprises and the potential risk connections;

[0184] A report generation unit is used to generate an intelligent diagnosis report, including: detailed information of the first-category node enterprise, enterprises related to the potential risk connection and the second-category node enterprise; and the potential chain break risk.

[0185] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0186] The advantages of the above technical solution are:

[0187] Through the multi-dimensional qualification evaluation model construction module, the system can comprehensively consider the industry characteristics and build a more scientific and reasonable evaluation model. This makes the qualification evaluation of enterprises more comprehensive and accurate, laying a solid foundation for the subsequent industrial chain construction.

[0188] The data acquisition and processing module acquires data from multiple sources and performs cleaning, verification and standardization. This ensures the comprehensiveness, accuracy and consistency of the data and improves the reliability of subsequent analysis.

[0189] The qualification analysis module uses a multi-dimensional qualification evaluation model to conduct quantitative analysis of enterprise qualifications. This method can objectively and accurately reflect the performance of enterprises in various dimensions, providing a reliable basis for enterprise classification and ranking.

[0190] The enterprise classification and sorting module achieves scientific classification and reasonable sorting of enterprises through cluster analysis and comprehensive index calculation. This helps to identify different types of enterprises and accurately sort them within each category, providing an important reference for the optimal configuration of the industrial chain.

[0191] The interactive data acquisition module collects data on supply, cooperation, competition, and technology dependence. This comprehensive interactive data helps to more accurately depict the complex relationship network between enterprises.

[0192] The intelligent industrial chain network diagram construction module realizes the intelligent construction of the industrial chain through intelligent matching algorithms, enterprise pairing and network diagram construction. This not only improves the construction efficiency, but also better reflects the actual relationship and interaction between enterprises.

[0193] The system can calculate the intensity of collaboration and resource flow between enterprises based on real-time data and integrate it into the network diagram. This enables the industrial chain network diagram to dynamically reflect changes in relationships between enterprises, improving the timeliness and accuracy of the network diagram.

[0194] The intelligent diagnosis module can comprehensively diagnose the health status of the industrial chain network and identify key node enterprises and potential risks through network analysis, node identification, risk assessment and other steps. This provides important support for the optimization and risk management of the industrial chain.

[0195] The intelligent industrial chain network diagram and intelligent diagnosis report generated by the system provide decision makers with an intuitive and comprehensive industrial chain view and optimization suggestions, which greatly improves the scientificity and efficiency of decision-making.

[0196] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the above methods is implemented.

[0197] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. Of course, there are other ways of readable storage media, such as quantum memory, graphene memory, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0198] The present invention also provides an electronic device. The electronic device of an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the intelligent industrial chain construction method based on multi-dimensional qualification certification provided by the present invention.

[0199] Example 10

[0200] Reference below Fig. 9 , which shows a schematic diagram of the structure of a computer system 300 of an electronic device suitable for implementing an embodiment of the present invention. Fig. 9 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0201] like Fig. 9As shown, the computer system 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage part 303 to a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the computer system 300 are also stored. The CPU 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0202] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0203] In particular, according to the embodiments disclosed in the present invention, the process described in the main step diagram above can be implemented as a computer software program. For example, the embodiments of the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the main step diagram. In the above embodiment, the computer program can be downloaded and installed from the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit 301, the above functions defined in the system of the present invention are executed.

[0204] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0205] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0206] The units involved in the embodiments of the present invention may be implemented by software or hardware. The units described may also be set in a processor. For example, it may be described as follows: a processor includes a pre-response unit, a receiving unit, and a request unit. The names of these units do not, in some cases, limit the units themselves.

[0207] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for constructing an intelligent industrial chain based on multi-dimensional qualification certification, characterized in that: The following steps are involved: S1: Construct a multi-dimensional qualification evaluation model based on industry characteristic data; S2: Obtain the pre-processed multi-dimensional qualification data of related enterprises in the industrial chain; S3: Analyze the multi-dimensional qualification data using the multi-dimensional qualification evaluation model to obtain multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain; S4: Based on the multi-dimensional quantitative indicators of the qualification data, a machine learning algorithm is used to classify the enterprises related to the industrial chain to obtain the enterprise category distribution, and the enterprises related to the industrial chain are comprehensively scored and ranked according to the set weights to obtain the ranking results; S5: Obtain interaction data, transaction intensity data, and resource flow data between related enterprises in the industrial chain; S6: Construct an intelligent industrial chain network diagram based on the multi-dimensional quantitative indicators, the enterprise category distribution, the ranking results, the industry characteristic data, the interaction data, the transaction intensity data and the resource flow data.

2. The method according to claim 1, characterized in that The S1 includes: S11: Obtain industry characteristic data including industrial policy information, industry standards, market demand information and intellectual property information; S12: Performing text analysis on the industry characteristic data using a natural language processing algorithm to extract key features and themes related to the industry; S13: Based on the expert knowledge base, the key features and the subject matter, determine the target dimensions of qualification assessment, and set assessment indicators and weights for each target dimension, wherein the weights reflect the relative importance of each dimension in industrial development; S14: Using machine learning algorithms, the evaluation indicators and weights of different target dimensions are normalized and weighted combined to form a unified multi-dimensional qualification assessment model.

3. The method according to claim 1, characterized in that The S3 includes: S31: inputting the multi-dimensional qualification data into the multi-dimensional qualification assessment model; S32: Using the multi-dimensional qualification evaluation model to process and calculate the input multi-dimensional qualification data, specifically including: For each target dimension, apply the corresponding evaluation indicators and weights to calculate the score of a single target dimension: Si = Σ(wij*xij); where Si is the score of the i-th target dimension, wij is the weight of the j-th indicator under the i-th target dimension, and xij is the standardized value of the j-th indicator under the i-th dimension; Calculate the overall qualification assessment score: S = Σ(Wi*Si), where S is the overall qualification assessment score and Wi is the weight of the i-th target dimension; Output multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain, wherein the multi-dimensional quantitative indicators include specific indicator values ​​of each target dimension and an overall qualification assessment score.

4. The method according to claim 1, characterized in that: The S4 includes: S41: Based on multi-dimensional quantitative indicators, the K-means clustering algorithm is used to perform cluster analysis on related enterprises in the industrial chain to obtain enterprise category distribution, which includes the number of enterprises in each category, category characteristics and relative relationships between categories; S42: According to the weight set for each target dimension, the multi-dimensional quantitative indicators are weighted averaged to obtain the comprehensive indicator of the enterprise. The calculation formula of the comprehensive qualification indicator is as follows: comprehensive qualification indicator = Σ(Wi*Si) / Σ(Wi), wherein Wi is the weight of the i-th target dimension, and Si is the score of the i-th target dimension; S43: Sort the enterprises in each category according to the comprehensive index to obtain a sorting result.

5. The method according to claim 1, characterized in that: The S6 includes: S61: Establishing an intelligent matching algorithm based on the multi-dimensional quantitative indicators, the enterprise category distribution, the ranking results and the industry characteristic data; S62: Optimizing the pairing of related enterprises in the industrial chain by using the intelligent matching algorithm to obtain a preliminary industrial chain collaboration relationship network; S63: Based on the preliminary industrial chain collaboration relationship network, a first industrial chain network diagram is constructed, wherein enterprise nodes represent enterprises in the industrial chain, relationship edges represent collaboration relationships between enterprises, and attributes of enterprise nodes include multi-dimensional quantitative indicators, enterprise category distribution, and ranking results; S64: Integrate the acquired interaction data between related enterprises in the industrial chain into the first industrial chain network graph, update the relationship edge type and relationship strength weight in the first industrial chain network graph, and obtain a second industrial chain network graph; S65: Calculate the collaboration intensity between enterprises based on the transaction intensity data, calculate the resource flow between enterprises based on the resource flow data, and add the collaboration intensity and the resource flow to the attribute information of the second industrial chain network diagram to form a final intelligent industrial chain network diagram.

6. The method according to claim 1, characterized in that The S61 specifically includes: S611: construct a feature vector space, and represent each enterprise as a multi-dimensional feature vector, wherein the multi-dimensional feature vector includes: the value of each indicator in the multi-dimensional quantitative indicator, the code of the category to which the enterprise belongs, and the ranking percentage of the enterprise in the category to which it belongs; S612: Determine a similarity calculation method based on the constructed feature vector space, and calculate similarity results between enterprises based on the similarity calculation method; S613: Using the multi-dimensional feature vector and the similarity results between the enterprises, a machine learning-based enterprise matching model is trained to predict the matching probability between enterprises.

7. The method according to claim 6, characterized in that The S612 specifically includes: Calculate the Euclidean distance between the multi-dimensional feature vectors to obtain a first similarity value; Using a predefined enterprise category compatibility matrix, based on the category code of the enterprise, determining compatibility weights between enterprises of different categories, and multiplying the compatibility weights by the first similarity value to obtain a second similarity value; Calculate a ranking influence factor according to the ranking percentage of the enterprise in the category to which it belongs, and use the ranking influence factor to multiply or add the second similarity value to obtain a third similarity value; The similarity result between the enterprises is obtained by weighted average calculation based on the first similarity value, the second similarity value and the third similarity value.

8. The method according to claim 6, characterized in that The S613 specifically includes: Choose a machine learning-based enterprise matching model from random forest or gradient boosted decision tree; The multidimensional feature vector and the similarity results between the enterprises are used as inputs of an enterprise matching model based on machine learning, and the enterprise matching model based on machine learning is trained so that it can predict the matching probability between enterprises.

9. The method according to claim 1, characterized in that: Also includes: S7: Based on the intelligent industrial chain network diagram, intelligent diagnosis is performed on the intelligent industrial chain network to generate an intelligent diagnosis report.

10. An intelligent industrial chain construction system based on multi-dimensional qualification certification, characterized in that: include: The qualification evaluation model construction module is used to build a multi-dimensional qualification evaluation model based on industry characteristic data; The data acquisition module is used to obtain the pre-processed multi-dimensional qualification data of related enterprises in the industrial chain; A qualification analysis module, used to analyze the multi-dimensional qualification data using the multi-dimensional qualification evaluation model to obtain multi-dimensional quantitative indicators of the qualification data of related enterprises in the industrial chain; An enterprise classification and ranking module is used to classify the enterprises related to the industrial chain based on the multi-dimensional quantitative indicators of the qualification data, adopt a machine learning algorithm to obtain the enterprise category distribution, and comprehensively score and rank the enterprises related to the industrial chain according to the set weights to obtain the ranking results; Interaction data acquisition module, used to obtain interaction data, transaction intensity data and resource flow data between related enterprises in the industrial chain; A network diagram construction module is used to construct an intelligent industrial chain network diagram based on the multi-dimensional quantitative indicators, the enterprise category distribution, the sorting results, the industry characteristic data, the interaction data, the transaction intensity data and the resource flow data.

Images