An industrial brain system based on big data and artificial intelligence and its implementation method

Abstract

This invention belongs to the technical field of industrial analysis systems, specifically an industrial brain system and its implementation method based on big data and artificial intelligence. This industrial brain system achieves multi-dimensional, in-depth analysis of the entire industrial chain through multi-model fusion. It can integrate heterogeneous industrial data from multiple sources, clearly present the overall development status of the industrial chain, accurately identify missing links in the industrial chain, and uncover potential development opportunities. It can provide reliable decision-making support for industrial planning, investment attraction, enterprise services, and policy formulation. Simultaneously, it enables precise information delivery to meet the needs of different stakeholders, effectively improving the matching degree and efficiency of industrial services, helping all parties involved in the industry efficiently obtain the resources they need, promoting the rational allocation of industrial resources, and contributing to the healthy development of the industrial ecosystem.

Description

Technical Field

[0001] This invention belongs to the field of industrial analysis system technology, specifically relating to an industrial brain system and its implementation method based on big data and artificial intelligence. Background Technology

[0002] Industry analysis is a crucial decision-making support for regional industrial development planning, risk warning, and policy formulation. Traditional industry analysis models mostly rely on manual surveys and experience-based judgments. Limited by data coverage and low data processing efficiency, they cannot efficiently integrate and aggregate cross-segment and cross-entity industrial operation data. Most analyses only stay at the macro-scale statistical level, lacking refined quantitative modeling of value distribution in the industrial chain, supply and demand matching, and the innovation capabilities of market entities, making it difficult to reveal the inherent laws of industrial operation. At the same time, existing systems are mostly single-function and cannot provide intelligent decision-making support for actual needs such as industrial management and investment attraction, failing to meet the actual needs of digital governance of industries in the context of the digital economy.

[0003] With the development of digital technology, big data and artificial intelligence technologies have been gradually applied to the field of industrial analysis. However, most existing industrial analysis systems focus on single-dimensional monitoring of industrial operations. They lack the ability to integrate and analyze the industrial chain, value chain, and supply and demand chain from multiple dimensions. They also struggle to dynamically predict industrial risks and iterate emergency plans, thus failing to provide systematic decision support for industrial governance. Summary of the Invention

[0004] This invention aims to provide an industrial brain system and its implementation method based on big data and artificial intelligence, so as to achieve integrated analysis of all-dimensional industrial operation data through the industrial brain system based on big data and artificial intelligence, construct a refined industrial quantitative analysis model, and provide intelligent decision support for scenarios such as industrial governance, investment promotion and talent attraction.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Provides an industry brain system based on big data and artificial intelligence, including: The data source layer includes an innovation entity database, an innovation resource database, a demand database, and an investment attraction database. The quality control layer performs operations including data cleaning, data deduplication, data association, data source classification, and metadata mapping. The knowledge base layer provides knowledge data including park knowledge, enterprise knowledge, talent knowledge, policy knowledge, intellectual property rights, and collaborative innovation competitions. It manages and applies the knowledge data through reasoning engines, storage technologies, efficient retrieval, and self-optimization mechanisms. The model layer includes an industrial cluster evolution model, an industrial panorama analysis model, a risk analysis model, a value chain analysis model, a supply and demand chain analysis model, an innovation capability assessment model, a multi-dimensional recommendation model, and a complex network model. The service layer provides services including industry data cloud, industry operation monitoring, industry panorama analysis, industrial chain risk analysis, industry value chain analysis, industry supply and demand chain analysis, enterprise innovation capability assessment, enterprise industrial chain cooperation analysis, enterprise investment and financing relationship analysis, potential investment attraction enterprise recommendation, potential talent recommendation, and potential policy recommendation.

[0006] Preferably, the industrial cluster evolution model analyzes the ecological synergy and symbiotic characteristics of the industrial cluster using the following formula (1): ; in, To integrate the importance of networks with environmental factors to ultimately determine the cluster correlation of enterprises; α represents the original PageRank network importance value of the i-th enterprise node; α is the environmental comprehensive correction parameter, which is influenced by external clustering factors including geographical location, resources, culture, and policies. It is the sum of the original PageRank network importance values ​​and the environmental comprehensive adjustment parameters for all enterprises in the entire network.

[0007] Preferably, the industry panorama analysis model includes an industry competitiveness evaluation index module and an industry spatial clustering analysis module, wherein: The industrial competitiveness evaluation index module calculates absolute indicators using the following formula (2), calculates relative indicators using the following formula (3), and performs dimensionless standardization on the index data using the following formula (4): Absolute quantity indicators: ; in, This represents the comprehensive absolute value of the h-th category indicator for region i and industry j after aggregation; i, j, and h are dimension subscripts, representing region, industry, and indicator category respectively; l represents the number of sub-basic indicators included under the h-th major category; w h The weight of the h-th sub-indicator; For region i and industry j, the original absolute value of the kth sub-indicator in the hth category; Relative quantity indicators: ; in, It is a relative ratio indicator used for horizontal comparative analysis between regions and industries; The original absolute value of the target research object; To compare the original absolute value of the benchmark object; Dimensionless standardization: ; Wherein, Standard_ind is the standard index value that takes the value in the interval [0,1] after dimensionless standardization; X is the original data value of the index; Xmax and Xmin are the maximum and minimum values ​​of the index in all samples; The industrial spatial agglomeration analysis module includes a regional difference analysis unit, a regional isomorphism analysis unit, a regional concentration index unit, and a spatial Gini coefficient unit. The regional difference analysis unit is used to analyze regional differences from two dimensions: region and industry. The regional isomorphism analysis unit is used to analyze the similarity of regional industrial structure development. The regional concentration index unit is used to calculate industrial concentration, that is, the share of the total output or other indicators of the largest industrial regions in the entire industry. The spatial Gini coefficient unit measures the degree of industrial spatial agglomeration by calculating the spatial Gini coefficient, and is used to measure the degree of industry agglomeration.

[0008] Preferably, the risk analysis model includes a supply chain risk early warning indicator module, a supply chain enterprise risk propagation analysis module, and a risk handling and contingency plan system update module, wherein: The industrial chain risk early warning indicator module is used to calculate the industrial risk score based on the industrial chain risk early warning indicator system and to conduct industrial chain risk early warning. The primary indicators of the industrial chain risk early warning indicator system include industrial strategicity, industrial pillar status, industrial foundation, industrial autonomy, and industrial leadership. The risk propagation analysis module for enterprises in the industrial chain uses the risk propagation dynamics equation (5) below to perform risk propagation analysis for enterprises in the industrial chain:

[0009]

[0010] The risk propagation analysis module of the industrial chain enterprises analyzes the industrial propagation range using the risk propagation range formula in the following equation (6):

[0011]

[0012] The risk propagation analysis module of the industrial chain enterprises analyzes the industrial propagation speed using the risk propagation speed formula in formula (7):

[0013] in, The average risk propagation speed after multiple simulations of risk source enterprise i; Let k be the total size of the enterprises affected by the risk in the simulation. This represents the time step consumed during the entire risk diffusion process in the k-th simulation.

[0014] The risk handling and contingency plan system update module is used to combine the results of the supply chain risk early warning indicator module and the supply chain risk propagation analysis module to analyze and evaluate supply chain risks. It integrates the risk contingency plan system and, based on accurate assessment, formulates tiered emergency response risk contingency plans for different alert states. It calculates the matching degree between risk events and risk contingency plans; the higher the value, the more suitable the current risk contingency plan is for handling the risk. At the same time, after risk handling, the risk contingency plan system is updated and iterated in a timely manner to achieve dynamic growth of the entire supply chain risk analysis model.

[0015] Preferably, the value chain analysis model calculates the total value ultimately created by the entire industry value chain using the following formula (8):

[0016] Where D represents the total value ultimately created by the entire industry value chain; D0 represents the initial basic value base of the value chain, that is, the inherent initial value and initial resource endowment of the industry chain, which is the core foundation for value creation; r G For the investment of enterprise G; r A Let D0 be the input of participant A, and D0 be the initial value before either party makes any input; k and v are the elasticity coefficients of input from both parties, respectively. The value chain analysis model calculates the company's expected revenue using the value incentive formula in equation (9):

[0017] The value incentive formula has the following constraints (9-1) and (9-2):

[0018] in, To maximize expected return; Let G and A be the expected revenue functions of the principal firm G and participant A, respectively; let x be the revenue share of the principal firm G and let 1-x be the revenue share of other participant A; IR is the compliance constraint, which states that the revenue of other participant A after making an investment cannot be less than a certain level. IC stands for compatible incentive, which is a constraint that ensures the consistency between the optimal strategy of the participants and the core enterprise's goals. Consistent with the parameter definition in formula (8); The value chain analysis model calculates the maximum expected revenue of the main enterprise G and other participants A using the following equations (10) and (11):

[0019] The parameters in equations (10) and (11) are consistent with the definitions above; among them, the exponents This is the comprehensive power of conversion under the equilibrium state of the game, used to balance the impact of the elasticity coefficients of both sides on the payoff; The value chain analysis model calculates the maximum overall total revenue from integrated value chain collaborative management using the following equation (12):

[0020] In this formula, the entire value chain belongs to a whole interest group, and the parameters are defined in the same way as the corresponding parameters in the previous formula. The value chain analysis model calculates the equilibrium benefits of the main enterprise G and other participants A when integrating value chain collaborative management using the following formulas (13) and (14):

[0021] in, The optimal equilibrium return for the main enterprise G; For the optimal equilibrium payoff of other participant A: the remaining parameters are the same as in the previous formula; The value chain analysis model calculates the optimal resource loss for a single link using the following formula (15):

[0022] Where, x ij F represents the resource input decision variable for the i-th link and j-th entity in the value chain, i.e., the scale of resource input by the entity in that link; t (x ij TR represents the net value output of the entity in this stage during period t, i.e., the difference between total revenue and total cost; t (x ij ) represents the total revenue of the entity in this stage during period t; TC t (x ij K represents the total cost of the main component in this stage during period t; r f(x) is the discount rate used to consider the time value of future income, discounting future income to its current value; t is the time period, usually in years; ij ) is the single-stage net value function corresponding to resource input, used to characterize the correspondence between resource input and net value output; The value chain analysis model calculates the maximum total value of the entire industry chain using the following formula (16):

[0023] Among them, V T V represents the total value of the entire industry value chain; m represents the total number of all links and participants in the value chain; j Let be the intrinsic value of the j-th value chain link.

[0024] Preferably, the supply and demand chain analysis model calculates the overall net revenue of the supply and demand chain using the following formula (17):

[0025] Where Y represents the total output and total revenue of the entire supply and demand chain system; I represents the total input cost of the entire supply and demand chain system; YI represents the overall net profit of the supply and demand chain, i.e., the total profit of the entire chain, used to characterize the overall profitability of the supply and demand chain; n represents the total number of all participating entities and sub-nodes within the supply and demand chain; r i The return on invested capital for the i-th participating entity is used to characterize the profitability of that entity's invested capital; i The scale of resource input by the i-th participating entity itself; The supply and demand chain analysis model calculates the supply and demand chain equilibrium measurement using the following formula (18):

[0026] in, Indicate production targets, C represents the actual production volume, k represents the market demand for the product, h represents the input-output ratio of resources to the product, and D represents the energy required in the production process. The supply and demand chain analysis model calculates the supply and demand chain resource allocation target at time t using the following formula (19):

[0027] Where f represents the supply and demand chain resource allocation target at time t; This indicates the amount of raw materials required for the product. Indicates the amount of recyclable raw materials in the product. Indicates the number of times a product is purchased by users. Indicates product output. This represents the remaining raw materials at time t. Indicates inventory cost; The supply and demand chain analysis model optimizes the balanced allocation of supply and demand chain resources using the following formula (20):

[0028] Where Z represents the optimized resource allocation result of the supply and demand chain, reflecting the degree of balance in resource allocation; W is the weighting coefficient, used to adjust the importance of different links or factors in resource allocation; U i This indicates the amount of resources available at each stage. This represents the average amount of resources obtainable at each stage. J This indicates the quantity of resources needed for supply. ρ represents resource utilization rate, and ρ represents resource demand rate.

[0029] Preferably, the innovation capability assessment model evaluates the enterprise's technological development level using the following formulas (21)-(24): Technology growth coefficient v t :

[0030] Technology maturity coefficient α t :

[0031] Technology aging coefficient β t :

[0032] New technology characteristic coefficient N t :

[0033] Where t is the t-th statistical year; a t b. The number of new invention patent applications in the target technology field in year t; t For year t, the number of new utility model patent applications in the target technology field; c t The number of new design patent applications in the target technology field in year t; The innovation capability assessment model uses the following formula (25) or (26) to predict the life cycle:

[0034] Where y(t) is the quantitative value of the level of technological development at time t, taking into account the cumulative number of patents and the scale of technological innovation output; t is time; K is the upper limit of technological development growth; a and b are constant parameters to be estimated in the model; and L is the ultimate upper limit of technological development growth.

[0035] Preferably, the multi-dimensional recommendation model integrates heterogeneous data from multiple sources by collecting content from data sources, then generates enterprise profiles for industry information mining, and performs demand analysis based on the information in the enterprise profiles to achieve recommendations for potential investment enterprises, potential talents, and potential policies from different demand perspectives.

[0036] Preferably, the complex network model includes an enterprise-industry chain complex network, used to reflect the overall relationship between enterprises and the industry chain. The enterprise-industry chain complex network is constructed as follows: a basic matrix is ​​constructed that reflects the relationship between various enterprises and industry chains. Each element in the basic matrix represents the relationship between the corresponding enterprise and industry chain. The basic matrix is ​​qualitatively processed to extract strong relationships, resulting in an industry association adjacency matrix. That is, by comparing the elements in the basic matrix, important relationships are marked as 1, and relatively unimportant relationships are marked as 0, thus obtaining an adjacency matrix that reflects whether there is a relationship between enterprises and industry chains. Based on the constructed adjacency matrix, the UCINET software is used to draw the relationship network diagram between enterprises and industry chains, thus obtaining the enterprise-industry chain complex network. The complex network model also includes a complex network of corporate investment and financing relationships. This complex network is established based on the equity fundraising or loan guarantee between companies. All companies are selected as the initial nodes of the complex network of investment and financing relationships, and the connections between nodes reflect the equity fundraising or loan guarantee relationships between companies.

[0037] The present invention also provides a method for implementing an industrial brain based on big data and artificial intelligence, wherein the method executes the functions of the industrial brain through the industrial brain system based on big data and artificial intelligence described in any of the preceding claims.

[0038] Compared with existing technologies, the beneficial effects of this invention are as follows: By constructing a multi-level, multi-dimensional model system encompassing industrial chain security analysis, value chain analysis, supply and demand chain analysis, innovation capability assessment, multi-dimensional recommendation, and complex network models, this invention covers the core links of industrial operation. It enables comprehensive and accurate perception of industrial development trends and risk warnings for key links, providing scientific data support for industrial management decisions, facilitating precise investment attraction and policy matching, and effectively improving the level of intelligent industrial governance. Each model in this invention has corresponding quantitative calculation formulas, with clearly defined parameters for different industrial dimensions. This transforms various abstract relationships and development states in industrial operation into quantifiable results, offering greater objectivity and reference value compared to qualitative analysis conclusions. Furthermore, it allows for dynamic updates and iterations to adapt to the ever-changing analytical needs of the industry. Attached Figure Description

[0039] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall architecture diagram of an embodiment of the industrial brain system based on big data and artificial intelligence of the present invention.

[0040] Figure 2 This is a diagram of the network architecture of an industrial cluster.

[0041] Figure 3 This is a flowchart of the supply chain risk analysis process of the risk analysis model in one embodiment of the industrial brain system based on big data and artificial intelligence of the present invention.

[0042] Figure 4 This is an architecture diagram of a multi-dimensional recommendation model in one embodiment of the industrial brain system based on big data and artificial intelligence of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] In one embodiment, an industry brain system based on big data and artificial intelligence is provided. This industry brain system revolves around the construction of the industrial chain and industrial ecosystem. Relying on new-generation information technologies such as cloud computing, big data, and artificial intelligence, as well as expert-assisted decision-making systems, it is designed for government departments, key enterprises, industrial parks, and investment institutions to build a full-industry chain data resource system. It realizes the collection of data across the entire industry chain, multi-source data storage, data quality control, and knowledge graph construction, providing support for data collaboration, data empowerment, and data application. It provides timely industrial data information and solutions to industrial problems for various entities in the industry chain, effectively guiding the entire industrial process, including R&D, production, management, marketing, and services, and accelerating the precise matching of resource elements and service needs.

[0045] like Figure 1 As shown, this industry brain system based on big data and artificial intelligence includes the following five layers: data source layer, quality control layer, knowledge base layer, model layer, and service layer.

[0046] The data source layer includes an innovation entity database, an innovation resource database, a demand database, and an investment attraction database. Organized according to the entire industry chain, the data source layer gathers multi-source heterogeneous resources such as regional industry data, industry sector data, online public opinion data, enterprise investment attraction data, industry chain cooperation data, and patents. By streamlining the knowledge resource system and data resource system, a full industry chain data resource center is constructed.

[0047] The operations performed by the quality control layer include data cleaning, data deduplication, data association, data source classification, and metadata mapping.

[0048] The quality control layer addresses issues such as data inconsistency, incompleteness, uncontrollability, and redundancy arising from multi-source data fusion through data cleaning, deduplication, data association, data source classification, and metadata mapping. Data cleaning identifies and corrects identifiable errors in data files, including consistency checks and handling invalid and missing values. Deduplication removes duplicate data from source data to ensure uniqueness. Data association extracts valuable relationships between data items from large datasets, reducing potentially massive amounts of disorganized data and transforming it into a smaller, easily observable and understandable document. Data source classification categorizes data sources based on characteristics such as origin, structure, variability, and volume, enhancing control over multiple data sources. Metadata mapping utilizes specific conversion programs to transform different metadata formats, facilitating resource management, organization, and utilization. Based on these multiple data quality control methods, the quality control layer ensures data quality at the data source layer, effectively promoting the construction of the knowledge base, model, and service layers of the industry brain system.

[0049] The knowledge base layer is the core of the industry brain, providing knowledge covering park knowledge, enterprise knowledge, talent knowledge, policy knowledge, intellectual property rights, and collaborative innovation competitions, and managing and applying knowledge data through inference engines, storage technologies, efficient retrieval, and self-optimization mechanisms.

[0050] The knowledge base layer is designed based on knowledge graphs. By processing large amounts of complex multimodal data, it forms a structured knowledge system that can be directly utilized. The knowledge base layer involves park knowledge, enterprise knowledge, talent knowledge, policy knowledge, intellectual property, and collaborative innovation competitions, etc. It relies on core technologies such as inference engines, storage technologies, efficient retrieval, self-optimization mechanisms, and reinforcement learning, and is built through six steps: knowledge modeling, knowledge extraction, knowledge fusion, knowledge storage, knowledge computation, and knowledge application. The modeling process is as follows: Knowledge modeling involves abstracting and modeling business problems according to the patterns agreed upon in the knowledge graph; knowledge extraction involves extracting knowledge from data of different sources and structures, forming structured data, and storing it in the knowledge graph; knowledge fusion involves merging descriptive information of the same entity or concept obtained from different sources; knowledge storage involves selecting an appropriate storage method based on business characteristics and data scale to save the fused data; knowledge computation involves discovering implicit relationships and knowledge in the structured knowledge repository; and knowledge application presents the "knowledge" built in the above steps to users in the most appropriate way, including combining various technologies to build semantic search, knowledge question answering, recommendation, and other systems. Several technical means were also used in the modeling process, such as inference engines, which rely on reinforcement learning technology to give knowledge graphs stronger reasoning, understanding and expression capabilities; the self-optimization mechanism is that the knowledge base adaptively adjusts parameters through identification technology and deep learning technology to ensure the self-expansion and self-optimization of the knowledge base.

[0051] The knowledge graph in the knowledge base layer can accurately represent the multi-level knowledge structure of different industry entities based on the data and resources generated throughout the entire life cycle of the industrial chain. On the one hand, it can handle complex multimodal industrial chain data, and on the other hand, it can provide multi-dimensional knowledge services such as industry, park, enterprise, talent, and policy, providing a foundation for modeling multi-industry knowledge.

[0052] The model layer includes industry cluster evolution models, industry panorama analysis models, risk analysis models, value chain analysis models, supply and demand chain analysis models, innovation capability assessment models, multi-dimensional recommendation models, and complex network models. This layer comprises business and data models built based on application scenarios and business collaborations to unlock the value of big data.

[0053] The model layer constructs multiple models to ensure the operation of the service layer, including: For the industrial cluster evolution model: combining Figure 2As shown, an industrial cluster is a geographically concentrated collection of interconnected companies and institutions within a specific field. As a unique form of industrial organization, industrial clusters effectively reduce transaction costs and improve economic efficiency through economies of scale and rational division of labor. Based on the logical framework for analyzing the formation of industrial clusters, this paper constructs economic models of synergistic coexistence, synergistic competition, and synergistic predation among enterprises within the cluster. An evolutionary model of the industrial cluster can be built using factors such as enterprise output, maximum output, and market saturation. This model analyzes the ecological synergy and symbiotic characteristics of industrial clusters, monitors industrial operations, and elucidates the ecological strategies for the survival and development of enterprises within the industrial cluster system, thus providing insights for formulating development strategies for industrial clusters and their constituent enterprises.

[0054] Industry Panorama Analysis Model: The industry panorama analysis model covers the industry cycle and the composition and transmission mechanism of the industrial chain. Based on the industry definition, it explores the industrial ecosystem construction model from upstream construction to downstream application, and draws a panoramic view of the industry. This allows for an objective and comprehensive evaluation of the industry's development status, reveals the difficulties and pain points in the industry's development, and provides a basis for macro-management decision-making.

[0055] Risk Analysis Model: In a globalized market competition environment, industrial development is susceptible to losses, damage, adverse effects, or even destruction due to factors such as technology, market conditions, and policies. The risk analysis model addresses the uncertainties in industrial development by centering on an industry early warning indicator system. It utilizes knowledge from multiple disciplines, including economics, management, statistics, and computer technology, and employs a combination of qualitative analysis by experts and quantitative data analysis to characterize, track, and forecast the development trends of industries. This aims to prevent harm from occurring unknowingly or when insufficiently prepared, and to minimize the losses caused by such harm.

[0056] Value Chain Analysis Model: From an industry perspective, competition between industries is essentially competition between industry value chains. The development of an industry largely depends on the advantages brought by its overall competitiveness. The Industry Brain empowers the value chain model, which uses a systematic approach to examine the activities and interrelationships throughout the entire lifecycle of each industry. It links upstream and downstream resources in the industrial chain, achieving precise resource matching. From a value chain perspective, it provides development strategies for enterprises in various industries in different directions—forward, backward, inward, and outward—to build regional core competitiveness and achieve business transformation and upgrading for enterprises and regions.

[0057] Supply and demand chain analysis model: A supply and demand chain is a network of supply and demand across multiple industries and entities, connected by logistics, information flow, and capital flow. The Industry Brain system examines the supply and demand network from the perspective of supply and demand chain optimization, combining policy, market development, and industry changes to improve the optimal decision-making challenges in supply and demand chain management caused by asymmetric information structures and inconsistent incentives.

[0058] Innovation Capability Assessment Model: Enterprise innovation is a key factor determining the direction, scale, and speed of enterprise development. To dynamically monitor and evaluate enterprise innovation activities and capabilities, and to provide reference for science and technology management and decision-making, the innovation capability assessment model constructs a series of evaluation indicators to accurately measure and reflect enterprise innovation capabilities and their role in transforming economic development patterns and implementing innovation-driven strategies. It selects different evaluation methods to comprehensively, objectively, and systematically reflect the overall capacity for continuous innovation within enterprises. The model comprehensively considers different levels of enterprise strategy, technology, knowledge, resource capabilities, organization, and efficiency to construct an evaluation indicator system. Combining various evaluation models such as the analytic hierarchy process (AHP) and comprehensive evaluation methods, it obtains enterprise innovation capability assessment results, providing a reference for governments to establish effective incentive mechanisms to promote enterprises' endogenous, continuous, and integrated innovation capabilities.

[0059] Multi-dimensional Recommendation Model: This model selects data from different dimensions for research, using mainstream recommendation algorithms such as collaborative filtering as its core. It addresses the problem of low recommendation quality caused by data sparsity and limited data in a single dimension. Simultaneously, it effectively combines similarity models and matrix factorization models, ensuring the reasonable application of both explicit and implicit feedback data, thereby reducing recommendation errors and improving accuracy. This model enables multi-dimensional industry monitoring and information dissemination, facilitating digital asset accumulation and assisting researchers in related industries with their research and analysis. Furthermore, by integrating different data sources, such as investment attraction databases and innovation entity databases, it provides services such as recommendations for potential investment enterprises, potential talents, and potential policies.

[0060] Complex network models, utilizing methods such as mathematical graph theory and statistical physics, study the relationships between elements and the structural characteristics of a system, exploring the evolutionary mechanisms and patterns of complex systems. An industrial chain is a complex network formed by the collaboration of multiple elements, entities, and connections, reflecting the connection methods between enterprises and institutions within a complex network. Industrial complex network models can effectively describe and deeply explain industrial relationships, structures, and effects, providing strong theoretical methods, feasible ideas, and strategic paths for further optimization of the relationship structure and enhancement of the effect level.

[0061] The service layer provides services including industry data cloud, industry operation monitoring, industry panorama analysis, industrial chain risk analysis, industrial value chain analysis, enterprise innovation capability assessment, enterprise industrial chain cooperation analysis, enterprise investment and financing relationship analysis, potential investment attraction enterprise recommendations, potential talent recommendations, and potential policy recommendations. The service layer is used for the precise matching of resource elements and service needs, providing a reference for industrial development.

[0062] Furthermore, the industrial cluster evolution model based on big data and artificial intelligence analyzes the ecological synergy and symbiotic characteristics of industrial clusters through the following formula (1):

[0063] in, To integrate the importance of networks with environmental factors to ultimately determine the cluster correlation of enterprises; α represents the original PageRank network importance value of the i-th enterprise node; α is the environmental comprehensive correction parameter, which is influenced by external clustering factors including geographical location, resources, culture, and policies. It is the sum of the original PageRank network importance values ​​and the environmental comprehensive adjustment parameters for all enterprises in the entire network.

[0064] In an industrial cluster, core enterprises are constantly developing. To adapt to this development, they need to establish extensive connections with other related enterprises within the cluster. New enterprises entering the cluster are also attracted by strong enterprises and establish connections with them. As policies influence enterprise factors for regulation, the probability of optimal connections between enterprise nodes also changes. Therefore, this model is based on the PageRank algorithm to establish a co-aggregation model to calculate the probability of a new node linking with an existing node i.

[0065] The builders of industrial clusters hope that the industrial clusters can always be in a stable upward phase, and can also upgrade the industrial structure when appropriate to strengthen their own survival and competitiveness, and ensure the positive and healthy development of the industrial clusters. Therefore, in the process of industrial cluster evolution, builders need to take certain measures to intervene and provide the cluster with appropriate systems and rules. This is conducive to inducing the positive factors of the cluster itself, mitigating the impact of the cluster's own unfavorable conditions, and promoting the healthy upward development of the cluster. Formula (1) can reflect the probability of a new enterprise node being linked to an existing enterprise node i, indicating the competitiveness of the industrial cluster and providing a reference for taking policy measures.

[0066] Furthermore, the industry panorama analysis model includes an industry competitiveness evaluation index module and an industry spatial clustering analysis module, wherein: The industrial competitiveness evaluation index module calculates absolute indicators using the following formula (2), calculates relative indicators using the following formula (3), and performs dimensionless standardization on the index data using the following formula (4): Absolute quantity indicators:

[0067] in, This represents the comprehensive absolute value of the h-th category indicator for region i and industry j after aggregation; i, j, and h are dimension subscripts, representing region, industry, and indicator category respectively; l represents the number of sub-basic indicators included under the h-th major category; w h The weight of the h-th sub-indicator; This represents the original absolute value of the k-th sub-indicator in the h-th category under region i and industry j.

[0068] In practical work, it is often difficult to draw objective and comprehensive conclusions by relying solely on absolute value analysis. Therefore, it is necessary to use it in combination with relative indicators.

[0069] Relative quantity indicators:

[0070] in, It is a relative ratio indicator used for horizontal comparative analysis between regions and industries; The original absolute value of the target research object; To compare with the original absolute value of the benchmark object.

[0071] Relative indicators demonstrate the composition of the whole and the quantitative relationships between its parts.

[0072] In a multi-indicator system, since the units and dimensions of each indicator are different, it is not convenient to analyze them. Therefore, it is necessary to standardize the indicators according to the general data processing method to make them standard values ​​between 0 and 1 before analysis and evaluation, so as to avoid the incomparability caused by the inconsistency of units between indicators. Therefore, dimensionless standardization processing is required.

[0073] Dimensionless standardization:

[0074] Wherein, Standard_ind is the standard index value in the interval [0,1] after dimensionless standardization; X is the original data value of the index; Xmax and Xmin are the maximum and minimum values ​​of the index in all samples.

[0075] The industrial competitiveness evaluation index module is based on a three-level industrial chain competitiveness evaluation index system: the criterion level, the element level, and the indicator level. The criterion level mainly includes scale index, technology index, system index, and development index. The element level evaluates the criteria based on different elements; for example, the scale index includes output, efficiency, and resources. The indicator level uses different indicators to reflect the content of the evaluated elements; for example, efficiency is reflected through major indicators such as profit and tax. The specific industrial competitiveness evaluation index system is shown in Table 1 below. Table 1 - Evaluation Index System for Industrial Chain Competitiveness

[0076] The industrial spatial agglomeration analysis module includes a regional difference analysis unit, a regional isomorphism analysis unit, a regional concentration index unit, and a spatial Gini coefficient unit. The regional difference analysis unit is used to analyze regional differences from two dimensions: region and industry. The regional isomorphism analysis unit is used to analyze the similarity of regional industrial structure development. The regional concentration index unit is used to calculate industrial concentration, that is, the share of the total output or other indicators of the largest industrial regions in the entire industry. The spatial Gini coefficient unit measures the degree of industrial spatial agglomeration by calculating the spatial Gini coefficient, and is used to measure the degree of industry agglomeration.

[0077] For the regional difference analysis unit, there are two analysis dimensions: region and industry. For example, the region dimension can be divided into eastern, central, western, northeastern and provinces, and the industry dimension can be divided into pharmaceutical manufacturing, aerospace manufacturing, electronic communication equipment manufacturing, electronic computer and office equipment manufacturing and medical equipment and instrument manufacturing, etc. The measures of difference analysis are divided into main business revenue, export delivery value, number of employees, number of enterprises, gross profit and net profit.

[0078] The regional isomorphism analysis unit is used to analyze regional industrial isomorphism. Industrial isomorphism can be measured using methods such as grey relational degree measurement, similarity coefficient measurement, and nonparametric test. Regional industrial isomorphism refers to the degree of structural similarity between regions that gradually emerges and intensifies during the process of industrial structure change in a region. Industrial differences and isomorphism are complementary concepts, both of which can measure similarity. Analyzing both is more conducive to resource allocation.

[0079] The factors influencing regional industrial differences and isomorphism are generally considered to be internal and external. Internal factors include two main categories: demand factors and supply factors. Supply is the foundation and prerequisite for industrial structure evolution, including natural resources, human resources, and technological resources. Demand mainly includes consumption demand and investment demand, which are highly correlated. The more similar the supply in a region, the stronger the industrial isomorphism. External factors include economic development strategies and policies. If a region can correctly understand and apply the laws of economic development and scientifically formulate strategies, its industries will continue to develop well. A correct development strategy is conducive to balancing the relationship between regional industrial differences and isomorphism.

[0080] The regional concentration index unit analyzes industry concentration. Industry concentration refers to the share of the total output or other indicators of the largest regions in the industry. It is the most commonly used and simplest absolute indicator that can measure the degree of competition in a certain industry.

[0081] Industry concentration can be reflected by the Herschman-Hirschman-Frandall index, also known as the H index, which refers to the sum of the squares of the market share of all firms in an industry. The index ranges from 0 to 1. The higher the value, the higher the degree of industry concentration. Its advantage is that it can accurately reflect the degree of industry concentration and promptly reflect changes in the degree of market monopoly and competition.

[0082] The spatial Gini coefficient, calculated using the spatial Gini coefficient unit, is used to measure the degree of industrial spatial agglomeration. It is used to assess the degree of industry agglomeration, and the formula is as follows:

[0083] G is the spatial Gini coefficient of industry, with a value of [0,1]. The larger the value, the higher the degree of spatial agglomeration; n is the total number of units in the study area; Si is the share of the scale of the industry in the i-th region in the total scale of the same industry in the country; Xi is the share of the overall economic scale of the i-th region in the overall national economy.

[0084] Furthermore, in combination Figure 3 As shown, the risk analysis model of the industry brain system centers on the industry risk early warning indicator system, integrates the industry risk propagation model, and adopts a combination of qualitative and quantitative analysis methods to track and analyze the development trend of the industry, predict early warnings, select contingency plans, and handle risks. The risk analysis model mainly includes three parts: the industrial chain risk early warning indicator module, the industrial chain enterprise risk propagation analysis module, and the risk handling and contingency plan system update module, among which: The industrial chain risk early warning indicator module is used to calculate the industrial risk score based on the industrial chain risk early warning indicator system and to conduct industrial chain risk early warning. The primary indicators of the industrial chain risk early warning indicator system include industrial strategicity, industrial pillar status, industrial foundation, industrial autonomy, and industrial leadership.

[0085] The industrial chain risk early warning indicator system adopted in this embodiment is shown in Table 2 below. The industrial chain risk early warning indicator system includes 5 primary indicators, 16 secondary indicators, and 38 tertiary indicators. Following the definition of tertiary indicators, the weight of each indicator is weighed using methods such as fuzzy evaluation and AHP analysis to calculate the final industrial risk score.

[0086] Table 2 - Industrial Chain Risk Early Warning Indicator System

[0087] The supply chain enterprise risk propagation analysis module uses the risk propagation dynamics equation (5) below to perform supply chain enterprise risk propagation analysis:

[0088]

[0089] The risk propagation analysis module for enterprises in the industrial chain analyzes the scope of industrial propagation using the risk propagation range formula in equation (6):

[0090]

[0091] The risk propagation analysis module for enterprises in the industrial chain analyzes the industry propagation speed using the risk propagation speed formula in equation (7):

[0092] in, The average risk propagation speed after multiple simulations of risk source enterprise i; Let k be the total size of the enterprises affected by the risk in the simulation. This represents the time step consumed during the entire risk diffusion process in the k-th simulation.

[0093] The risk management and contingency plan system update module combines the results of the supply chain risk early warning indicator module and the supply chain risk propagation analysis module to analyze and assess supply chain risks. It integrates the risk contingency plan system and, based on accurate assessment, formulates tiered emergency response risk contingency plans for different alert states. It calculates the matching degree between risk events and risk contingency plans; the higher the value, the more suitable the current risk contingency plan is for handling the risk. At the same time, after risk management, the risk contingency plan system is updated and iterated in a timely manner to achieve dynamic growth of the entire supply chain risk analysis model.

[0094] By establishing a supply chain risk analysis model, timely observation, analysis, and handling of risks can effectively mitigate the negative impact of risks. Furthermore, by integrating risk early warning indicator systems with risk propagation models, the likelihood of supply chain risk propagation and its widespread dissemination can be reduced from the source.

[0095] Furthermore, the value chain analysis model calculates the total value ultimately created by the entire industry value chain using the following general formula for value creation (8):

[0096] Where D represents the total value ultimately created by the entire industry value chain; D0 represents the initial basic value base of the value chain, that is, the inherent initial value and initial resource endowment of the industry chain, which is the core foundation for value creation; r G For the investment of enterprise G; r A Let D0 be the input of participant A, and D0 be the initial value before either party makes any input; k and v are the elasticity coefficients of input from both parties, respectively. The value chain analysis model calculates the firm's expected revenue using the value incentive formula in equation (9):

[0097] The value incentive formula has the following constraints (9-1) and (9-2):

[0098] in, To maximize expected return; Let G and A be the expected revenue functions of the principal firm G and participant A, respectively; let x be the revenue share of the principal firm G and let 1-x be the revenue share of other participant A; IR is the compliance constraint, which states that the revenue of other participant A after making an investment cannot be less than a certain level. IC stands for compatible incentive, which is a constraint that ensures the consistency between the optimal strategy of the participants and the core enterprise's goals. Consistent with the parameter definition in formula (8); The value chain analysis model calculates the maximum expected revenue of the main enterprise G and other participants A using the following equations (10) and (11):

[0099] The parameters in equations (10) and (11) are consistent with the definitions above; among them, the exponents This is the comprehensive power of conversion under the equilibrium state of the game, used to balance the impact of the elasticity coefficients of both sides on the payoff; The value chain analysis model calculates the maximum overall total revenue from integrated value chain collaborative management using the following equation (12):

[0100] In this formula, the entire value chain belongs to a whole interest group, and the parameters are defined in the same way as the corresponding parameters in the previous formula. The value chain analysis model calculates the equilibrium benefits of the main enterprise G and other participants A when integrating value chain collaborative management using the following formulas (13) and (14):

[0101] in, The optimal equilibrium return for the main enterprise G; For the optimal equilibrium payoff of other participant A: the remaining parameters are the same as in the previous formula; The value chain analysis model calculates the optimal resource loss for a single link using the following formula (15):

[0102] Where, x ij F represents the resource input decision variable for the i-th link and j-th entity in the value chain, i.e., the scale of resource input by the entity in that link; t (x ijTR represents the net value output of the entity in this stage during period t, i.e., the difference between total revenue and total cost; t (x ij ) represents the total revenue of the entity in this stage during period t; TC t (x ij K represents the total cost of the main component in this stage during period t; r f(x) is the discount rate used to consider the time value of future income, discounting future income to its current value; t is the time period, usually in years; ij ) is the single-stage net value function corresponding to resource input, used to characterize the correspondence between resource input and net value output; The value chain analysis model calculates the maximum total value of the entire industry chain using the following formula (16):

[0103] Among them, V T V represents the total value of the entire industry value chain; m represents the total number of all links and participants in the value chain; j Let be the intrinsic value of the j-th value chain link.

[0104] Furthermore, the supply and demand chain analysis model calculates the overall net revenue of the supply and demand chain using the following formula (17):

[0105] Where Y represents the total output and total revenue of the entire supply and demand chain system; I represents the total input cost of the entire supply and demand chain system; YI represents the overall net profit of the supply and demand chain, i.e., the total profit of the entire chain, used to characterize the overall profitability of the supply and demand chain; n represents the total number of all participating entities and sub-nodes within the supply and demand chain; r i The return on invested capital for the i-th participating entity is used to characterize the profitability of that entity's invested capital; i The scale of resource input by the i-th participating entity itself; The supply and demand chain analysis model calculates the supply and demand chain equilibrium measure using the following equation (18):

[0106] in, Indicate production targets, C represents the actual production volume, k represents the market demand for the product, h represents the input-output ratio of resources to the product, and D represents the energy required in the production process. The supply and demand chain analysis model calculates the supply and demand chain resource allocation target at time t using the following formula (19):

[0107] Where f represents the supply and demand chain resource allocation target at time t; This indicates the amount of raw materials required for the product. Indicates the amount of recyclable raw materials in the product. Indicates the number of times a product is purchased by users. Indicates product output. This represents the remaining raw materials at time t. Indicates inventory cost; The supply and demand chain analysis model optimizes the balanced allocation of supply and demand chain resources using the following formula (20):

[0108] Where Z represents the optimized resource allocation result of the supply and demand chain, reflecting the degree of balance in resource allocation; W is the weighting coefficient, used to adjust the importance of different links or factors in resource allocation; U i This indicates the amount of resources available at each stage. This represents the average amount of resources obtainable at each stage. J This indicates the quantity of resources needed for supply. ρ represents resource utilization rate, and ρ represents resource demand rate.

[0109] Furthermore, the innovation capability assessment model evaluates the enterprise's technological development level using the following formulas (21)-(24): Technology growth coefficient v t :

[0110] Technology maturity coefficient α t :

[0111] Technology aging coefficient β t :

[0112] New technology characteristic coefficient N t :

[0113] Where t is the t-th statistical year; a t b. The number of new invention patent applications in the target technology field in year t; t For year t, the number of new utility model patent applications in the target technology field; c t The number of new design patent applications in the target technology field in year t; The innovation capability assessment model uses the following formula (25) or (26) to predict the life cycle:

[0114] Where y(t) is the quantitative value of the level of technological development at time t, taking into account the cumulative number of patents and the scale of technological innovation output; t is time; K is the upper limit of technological development growth; a and b are constant parameters to be estimated in the model; and L is the ultimate upper limit of technological development growth.

[0115] Furthermore, the multi-dimensional recommendation model integrates heterogeneous data from multiple sources by collecting content from various data sources, then generates enterprise profiles for industry information mining, and performs demand analysis based on the information in the enterprise profiles to achieve recommendations for potential investment enterprises, potential talents, and potential policies from different demand perspectives.

[0116] Combination Figure 4 As shown, the multi-dimensional recommendation model consists of seven stages: data source layer, content collection, integration of multi-source heterogeneous data, enterprise profile generation, industry information mining, demand analysis, and function implementation.

[0117] The data source layer revolves around the construction of the industrial chain and industrial ecosystem, realizing the construction of databases such as the innovation entity database, innovation resource database, demand database, and investment attraction database. Simultaneously, due to the dispersed distribution of industry information resources, it is necessary to collect the required information from multiple sources, including industry websites, academic databases, government websites, corporate websites, industry organizations, library and information science institutions, and industry big data companies. This provides a continuous stream of fresh data resources for the data analysis applications of multi-dimensional recommendation models, realizing the underlying data infrastructure deployment of the industry brain.

[0118] For the content collection stage, data should be selected from the data source layer and comprehensively cover industry technical information, market information, product information, policy information, regulatory information, as well as basic enterprise information, intellectual property information, business-related data, customer comments, and recent developments.

[0119] In the process of integrating multi-source heterogeneous data, since information resources obtained from multiple channels may overlap and exhibit heterogeneity, it is necessary to standardize and integrate the data after collection to improve resource availability.

[0120] In the process of generating enterprise profiles, in order to better support the development of precise industry information services, enterprise profiles need to be able to comprehensively and accurately reflect the information needs of enterprise users. At the same time, adhering to the principle of simplicity, enterprise characteristics that cannot help analyze information needs should be avoided as much as possible in the profile system.

[0121] Needs analysis: The information needs of enterprises mainly include economic information, social information, policy information, market information, technology information, talent information, competitor information, supplier information, and investment information.

[0122] Functionality Implementation: In the implementation of the industry information precision service function based on enterprise profiles, the main value of enterprise profiles lies in enabling precise analysis of enterprise user needs. For example, in policy information demand analysis, it is necessary to accurately locate the required policy information by utilizing features such as industry, registered location, office location, and customer regional distribution. In market information analysis, it is necessary to further clarify the market information needs by using profile features such as product / service, customer regional characteristics, age characteristics, gender characteristics, and occupational characteristics. It can also rely on similarity models, knowledge graph association graph mining, and other models to achieve recommendations for potential investment enterprises, potential talents, and potential policies from different demand perspectives.

[0123] Furthermore, the complex network model includes the enterprise-industry chain complex network, which reflects the overall relationship between enterprises and the industry chain. The construction method of the enterprise-industry chain complex network is as follows: By constructing a basic matrix that reflects the relationship between various enterprises and industry chains, each element in the basic matrix represents the relationship between the corresponding enterprise and industry chain. The basic matrix is ​​qualitatively processed to extract strong relationships, resulting in an industry association adjacency matrix. That is, by comparing the elements in the basic matrix, important relationships are marked as 1, and relatively unimportant relationships are marked as 0, thus obtaining an adjacency matrix that reflects whether there is a relationship between enterprises and industry chains. Based on the constructed adjacency matrix, the UCINET software is used to draw the relationship network diagram between enterprises and industry chains, thus obtaining the enterprise-industry chain complex network. The complex network model also includes a complex network of corporate investment and financing relationships. This complex network is established based on the equity financing or loan guarantee relationships between companies. All companies are selected as the initial nodes of the complex network, and the connections between nodes reflect the equity financing or loan guarantee relationships between companies.

[0124] As demonstrated by the above embodiments, this industry brain system based on big data and artificial intelligence achieves multi-dimensional and in-depth analysis of the entire industry chain through multi-model fusion. It can integrate heterogeneous industry data from multiple sources, clearly present the overall development status of the industry chain, accurately identify missing links in the industry chain, and uncover potential development opportunities. This provides reliable decision-making support for industry planning, investment attraction, enterprise services, and policy formulation. Simultaneously, it enables precise information delivery to meet the needs of different stakeholders, effectively improving the matching degree and efficiency of industry services, helping all parties involved in the industry efficiently obtain the resources they need, promoting the rational allocation of industry resources, and contributing to the healthy development of the industry ecosystem. The innovation capability assessment module helps enterprises and management departments clearly grasp the technological development trends and current stages in their fields, make advance technology layout plans, avoid the risks of technological iteration, and seize the window of opportunity for innovative development. The complex network model can intuitively present the relationships between enterprises and within the industry chain, quickly identify core leading enterprises and key nodes, and analyze the structural characteristics of industrial development, providing precise direction for strengthening and supplementing the industry chain.

[0125] In one embodiment, a method for implementing an industry brain based on big data and artificial intelligence is provided, characterized in that: the method executes the functions of the industry brain through the industry brain system based on big data and artificial intelligence in the previous embodiment, and the specific functions include: (1) monitoring the operation status of the industry, based on the above-mentioned value chain analysis model and supply and demand chain analysis model, dynamically monitoring the operation status of the entire industry chain in the region, intuitively presenting the overall scale of the industry chain, revenue distribution, value ratio of links and supply and demand matching balance, and timely capturing abnormal fluctuations and structural imbalances in the operation of the industry; (2) analyzing the industry innovation capability, relying on the innovation capability assessment model, quantitatively assessing the technological innovation development level of key industries and key enterprises in the region, predicting the evolution trend of the technology life cycle, and clarifying the regional industry innovation. (3) Precise investment attraction recommendation: Based on the multi-dimensional recommendation model, combined with the regional industrial chain development plan and the existing industrial chain gap, match and screen potential investment target enterprises that meet the development positioning, and provide directional support for precise investment attraction; (4) Talent demand and policy matching recommendation: Based on the multi-dimensional recommendation model, combined with the industrial development direction and enterprise innovation development needs, match high-level potential talents that meet the needs, and at the same time, combined with the industry, development stage and core needs of the enterprise, accurately match suitable industrial support policies; (5) Industrial network structure analysis: Relying on the complex network model, analyze the overall network connection characteristics of regional industries, identify the core leading enterprises of the industrial chain, sort out the upstream and downstream connections of the industrial chain, find the key weak nodes in the industrial chain network, and provide decision-making basis for the industrial chain to be strengthened and supplemented.

[0126] It should be noted that, in this document, terms such as “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0127] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An industrial brain system based on big data and artificial intelligence, characterized in that, include: The data source layer includes an innovation entity database, an innovation resource database, a demand database, and an investment attraction database. The quality control layer performs operations including data cleaning, data deduplication, data association, data source classification, and metadata mapping. The knowledge base layer provides knowledge data including park knowledge, enterprise knowledge, talent knowledge, policy knowledge, intellectual property rights, and collaborative innovation competitions. It manages and applies the knowledge data through reasoning engines, storage technologies, efficient retrieval, and self-optimization mechanisms. The model layer includes an industrial cluster evolution model, an industrial panorama analysis model, a risk analysis model, a value chain analysis model, a supply and demand chain analysis model, an innovation capability assessment model, a multi-dimensional recommendation model, and a complex network model. The service layer provides services including industry data cloud, industry operation monitoring, industry panorama analysis, industrial chain risk analysis, industry value chain analysis, industry supply and demand chain analysis, enterprise innovation capability assessment, enterprise industrial chain cooperation analysis, enterprise investment and financing relationship analysis, potential investment attraction enterprise recommendation, potential talent recommendation, and potential policy recommendation.

2. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The industrial cluster evolution model analyzes the ecological synergy and symbiotic characteristics of industrial clusters through the following equation (1): ； in, To integrate the importance of networks with environmental factors to ultimately determine the cluster correlation of enterprises; α represents the original PageRank network importance value of the i-th enterprise node; α is the environmental comprehensive correction parameter, which is influenced by external clustering factors including geographical location, resources, culture, and policies. It is the sum of the original PageRank network importance values ​​and the environmental comprehensive adjustment parameters for all enterprises in the entire network.

3. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The industry panorama analysis model includes an industry competitiveness evaluation index module and an industry spatial clustering analysis module, wherein: The industrial competitiveness evaluation index module calculates absolute indicators using the following formula (2), calculates relative indicators using the following formula (3), and performs dimensionless standardization on the index data using the following formula (4): Absolute quantity indicators: ； in, This represents the comprehensive absolute value of the h-th category indicator for region i and industry j after aggregation; i, j, and h are dimension subscripts, representing region, industry, and indicator category respectively; l represents the number of sub-basic indicators included under the h-th major category; w h The weight of the h-th sub-indicator; For region i and industry j, the original absolute value of the kth sub-indicator in the hth category; Relative quantity indicators: ； in, It is a relative ratio indicator used for horizontal comparative analysis between regions and industries; The original absolute value of the target research object; To compare the original absolute value of the benchmark object; Dimensionless standardization: ； Wherein, Standard_ind is the standard index value that takes the value in the interval [0,1] after dimensionless standardization; X is the original data value of the index; Xmax and Xmin are the maximum and minimum values ​​of the index in all samples; The industrial spatial agglomeration analysis module includes a regional difference analysis unit, a regional isomorphism analysis unit, a regional concentration index unit, and a spatial Gini coefficient unit. The regional difference analysis unit is used to analyze regional differences from two dimensions: region and industry. The regional isomorphism analysis unit is used to analyze the similarity of regional industrial structure development. The regional concentration index unit is used to calculate industrial concentration, that is, the share of the total output or other indicators of the largest industrial regions in the entire industry. The spatial Gini coefficient unit measures the degree of industrial spatial agglomeration by calculating the spatial Gini coefficient, and is used to measure the degree of industry agglomeration.

4. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The risk analysis model includes a supply chain risk early warning indicator module, a supply chain enterprise risk propagation analysis module, and a risk handling and contingency plan system update module, wherein: The industrial chain risk early warning indicator module is used to calculate the industrial risk score based on the industrial chain risk early warning indicator system and to conduct industrial chain risk early warning. The primary indicators of the industrial chain risk early warning indicator system include industrial strategicity, industrial pillar status, industrial foundation, industrial autonomy, and industrial leadership. The risk propagation analysis module for enterprises in the industrial chain uses the risk propagation dynamics equation (5) below to perform risk propagation analysis for enterprises in the industrial chain: ； ； The risk propagation analysis module of the industrial chain enterprises analyzes the industrial propagation range using the risk propagation range formula in the following equation (6): ； ； The risk propagation analysis module of the industrial chain enterprises analyzes the industrial propagation speed using the risk propagation speed formula in formula (7): ； in, The average risk propagation speed after multiple simulations of risk source enterprise i; Let k be the total size of the enterprises affected by the risk in the simulation. This represents the time step consumed during the entire risk diffusion process in the k-th simulation. The risk handling and contingency plan system update module is used to formulate graded emergency response risk plans for different alert states, calculate the matching degree between risk events and risk plans, and update and iterate the risk plan system.

5. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The value chain analysis model calculates the total value created by the entire industry value chain using the following formula (8): ； Wherein, D is the total value of the entire industry value chain ultimately created; D0 is the initial value of the value chain foundation value base; r G is the input of enterprise G; r A is the input of participant A, D0 is the initial value of both parties without input; k and v are the elasticity coefficients of both parties' input; The value chain analysis model calculates the company's expected revenue using the value incentive formula in equation (9): ； The value incentive formula has the following constraints (9-1) and (9-2): ； in, To maximize expected return; Let G and A be the expected revenue functions of the principal firm G and participant A, respectively; let x be the revenue share of the principal firm G and let 1-x be the revenue share of other participant A; IR is the compliance constraint, which states that the revenue of other participant A after making an investment cannot be less than a certain level. IC stands for compatible incentive, which is a constraint that ensures the consistency between the optimal strategy of the participants and the core enterprise's goals. Consistent with the parameter definition in formula (8); The value chain analysis model calculates the maximum expected revenue of the main enterprise G and other participants A using the following equations (10) and (11): ； The parameters in equations (10) and (11) are consistent with the definitions above; among them, the exponents This is the comprehensive power of conversion under the equilibrium state of the game, used to balance the impact of the elasticity coefficients of both sides on the payoff; The value chain analysis model calculates the maximum overall total revenue from integrated value chain collaborative management using the following equation (12): ； In this formula, the entire value chain belongs to a whole interest group, and the parameters are defined in the same way as the corresponding parameters in the previous formula. The value chain analysis model calculates the equilibrium benefits of the main enterprise G and other participants A when integrating value chain collaborative management using the following formulas (13) and (14): ； in, The optimal equilibrium return for the main enterprise G; For the optimal equilibrium payoff of other participant A: the remaining parameters are the same as in the previous formula; The value chain analysis model calculates the optimal resource loss for a single link using the following formula (15): ； wherein x ij is the resource input decision variable of the i-th link of the value chain and the j-th subject; F t (x ij ) is the net value output of the subject of the link in the t-th period; TR t (x ij ) is the total income of the subject of the link in the t-th period; TC t (x ij ) is the total cost of the subject of the link in the t-th period; K r is the capital discount rate; t is the time period, usually in years; f(x ij ) is the single-link net value function corresponding to the resource input. The value chain analysis model calculates the maximum total value of the entire industry chain using the following formula (16): ； Wherein, V T is the total value of the entire industry value chain; m is the total number of all links and participants in the value chain; V j is the value of the jth link of the value chain.

6. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The supply and demand chain analysis model calculates the overall net revenue of the supply and demand chain using the following formula (17): ； Wherein, Y is the total output and total income of the supply and demand chain system; I is the total input cost of the supply and demand chain system; Y-I is the total net income of the supply and demand chain; n is the total number of all participating subjects and sub-nodes in the supply and demand chain; r i The input capital return rate of the i th participating subject i The resource input scale of the i th participating subject itself The supply and demand chain analysis model calculates the supply and demand chain equilibrium measurement using the following formula (18): ； in, Indicate production targets, C represents the actual production volume, k represents the market demand for the product, h represents the input-output ratio of resources to the product, and D represents the energy required in the production process. The supply and demand chain analysis model calculates the supply and demand chain resource allocation target at time t using the following formula (19): ； Where f represents the supply and demand chain resource allocation target at time t; This indicates the amount of raw materials required for the product. Indicates the amount of recyclable raw materials in the product. Indicates the number of times a product is purchased by users. Indicates product output. This represents the remaining raw materials at time t. Indicates inventory cost; The supply and demand chain analysis model optimizes the balanced allocation of supply and demand chain resources using the following formula (20): ； Where Z represents the optimized resource allocation result of the supply and demand chain; W represents the weighting coefficient; U i This indicates the amount of resources available at each stage. This represents the average amount of resources obtainable at each stage. J This indicates the quantity of resources needed for supply. ρ represents resource utilization rate, and ρ represents resource demand rate.

7. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The innovation capability assessment model evaluates the enterprise's technological development level using the following formulas (21)-(24): Technology growth coefficient v t : ; Technology maturity coefficient α t : ; Technology aging coefficient β t : ; New technology characteristic coefficient N t : ; Where t is the t-th statistical year; a t b. The number of new invention patent applications in the target technology field in year t; t For year t, the number of new utility model patent applications in the target technology field; c t The number of new design patent applications in the target technology field in year t; The innovation capability assessment model uses the following formula (25) or (26) to predict the life cycle: ； Where y(t) is the quantitative value of the level of technological development at time t; t is time; K is the upper limit of technological development growth; a and b are constant parameters to be estimated in the model; and L is the ultimate upper limit of technological development growth.

8. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The multi-dimensional recommendation model integrates heterogeneous data from multiple sources by collecting content from data sources, then generates enterprise profiles for industry information mining, and performs demand analysis based on the information in the enterprise profiles to achieve recommendations for potential investment enterprises, potential talents, and potential policies from different demand perspectives.

9. The industrial brain system based on big data and artificial intelligence according to claim 1, characterized in that: The complex network model includes a firm-industry chain complex network, which reflects the overall relationship between firms and the industry chain; the complex network model also includes a firm investment and financing relationship complex network, which is established based on firms raising funds through equity financing or loan guarantees.

10. A method for implementing an industrial brain based on big data and artificial intelligence, characterized in that: The method performs the functions of the industrial brain through the industrial brain system based on big data and artificial intelligence as described in any one of claims 1-9.

Images