An analysis method, system and related equipment for regional industrial network evolution

Abstract

The present invention discloses an analysis method, system and related equipment for regional industrial network evolution, the analysis method includes: constructing an enterprise information database for specific regional industrial network evolution analysis; based on the enterprise information database, sequentially extracting enterprise growth time series data of different industries in the region, constructing a regional industry-specific growth trend prediction model to predict the future growth of enterprises in different industries in the specific region; constructing an enterprise relationship network based on the enterprise information database, removing enterprises with high extinction risk, predicting the potential relationship between the remaining nodes based on the nodes and relationships of the remaining industrial network through a link prediction method, and realizing industrial network evolution prediction in combination with the growth of enterprises in different industries. The use of time series similarity to improve the industrial development trend prediction model can effectively reduce the error of short-term prediction and ensure the prediction of long-term growth trend; using enterprise-related data of multiple dimensions, the analysis results are more scientific.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and more specifically to an analysis method, system and related equipment for the evolution of regional industrial networks. Background Art

[0002] At present, the existing network structure level and network structure characteristic analysis of industrial clusters are static descriptions of industrial cluster networks. As a social and economic network system, the formation and development of industrial clusters are always in a dynamic evolution process. The network evolution mechanism of industrial clusters is not only the driving force for the continuous development and growth of industrial cluster networks, but also the root cause of industrial cluster risks. Therefore, under the current downward trend of the overall economic situation, it is necessary to form a method for analyzing the evolution of industrial networks, analyze and predict the growth and extinction trends of industrial network nodes and edges, which is conducive to the early perception of the risks faced by the evolution of network structures, and provide decision-making support for targeted guidance work and optimization of the business environment.

[0003] At present, existing analysis methods are more inclined to current situation analysis, and the analysis of future trends is more inclined to qualitative judgment. It lacks quantitative support and has insufficient dynamic analysis capabilities for the evolution trend of industrial networks. This leads to limited support for the decision-making of relevant research results for regional industrial management departments. Summary of the invention

[0004] In this embodiment, a method, system, electronic device and storage medium for analyzing the evolution of a regional industrial network are provided to solve the problem of insufficient dynamic analysis capability of the evolution trend of the industrial network in related technologies.

[0005] In a first aspect, an embodiment of the present invention provides an analysis method for regional industrial network evolution, the analysis method comprising:

[0006] Obtain enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for specific regional industrial network evolution analysis;

[0007] Based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-specific growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region;

[0008] Based on the enterprise information database, construct enterprise feature vectors and construct positive and negative sample data sets from two dimensions: enterprise business activity characteristics and external environment characteristics; divide the positive and negative sample data sets into training sets and test sets; use the training set and test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the risk of enterprise extinction;

[0009] Based on the enterprise information database, an enterprise relationship network is constructed to remove enterprises with high extinction risks. Based on the nodes and relationships of the remaining industrial network, the potential relationships between the remaining nodes are predicted through link prediction methods, and the industrial network evolution prediction is realized in combination with the growth of enterprises in different industries.

[0010] Optionally, obtain the enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for specific regional industrial network evolution analysis, including:

[0011] Based on the enterprise registration data, the unified credit code is used as the unique identifier to clean and pre-process the registration data to form an enterprise registration information database;

[0012] According to the unified credit code of the enterprise, the enterprise data is aligned and integrated to obtain the enterprise business activity database;

[0013] Obtain open source data from the Internet, including regional POI data, population density data, transportation facility data, office leasing data, statistical yearbook data, and corporate litigation data, and obtain corporate external environment data through geocoding based on the address information contained in the open source data;

[0014] Based on the enterprise registration information database, enterprise business activity database and enterprise external environment data, an enterprise information database for specific regional industrial network evolution analysis is constructed.

[0015] Optionally, based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-by-industry growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region, including:

[0016] Based on the enterprise information database, combined with the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn. ;

[0017] Construct a regional and industry-specific growth trend forecasting model based on enterprise growth time series data;

[0018] The numerical solution of the classical growth model is obtained by the differential equation numerical method, and the initial analysis results are obtained. ;

[0019] Combining initial analysis results Establish Timing analysis model of

[0020] The parameters of the time series analysis model are obtained through the time series parameter estimation method;

[0021] Use the time series analysis model to predict the future growth of enterprises in different industries in a specific region.

[0022] Optionally, based on the enterprise information database, construct enterprise feature vectors and construct positive and negative sample data sets from two dimensions: enterprise business activity characteristics and external environment characteristics; divide the positive and negative sample data sets into training sets and test sets; use the training set and the test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the risk of enterprise extinction, including:

[0023] Based on the enterprise information database, extracting enterprise business activity characteristics and external environment characteristics, wherein the enterprise business activity characteristics include static characteristics and dynamic characteristics;

[0024] Construct positive and negative sample sets based on the business status of the enterprise, and divide the enterprise into training set and test set by random sampling;

[0025] The XGBOOST model and LightGBM model are trained using the training set and test set respectively;

[0026] Based on the trained XGBOOST model and LightGBM model, the enterprise extinction risk prediction is realized. The enterprise extinction risk prediction results include high extinction risk and low extinction risk.

[0027] Optionally, the risk of enterprise extinction is predicted based on the trained XGBOOST model and LightGBM model. The calculation method is as follows:

[0028] ;

[0029] Among them, Pi represents the extinction probability of enterprise i, which is the average of the predicted probabilities of the XGBOOST model and the LightGBM model.

[0030] Among them, Pi greater than or equal to 0.5 is considered a high extinction risk, and Pi less than 0.5 is considered a low extinction risk.

[0031] Optionally, an enterprise relationship network is constructed based on the enterprise information database, enterprises with high extinction risk are removed, and based on the nodes and relationships of the remaining industrial network, potential relationships between the remaining nodes are predicted by a link prediction method, and the industrial network evolution prediction is realized in combination with the growth of enterprises in different industries, including:

[0032] Extract enterprise relationships in sequence, and use the entropy method to calculate the relationship weights of enterprises to form an enterprise relationship network. Enterprise relationships include investment relationships, supply chain relationships, industrial chain relationships, and cooperative relationships.

[0033] Eliminate enterprises with high extinction risk and their connection edges, and use link prediction methods based on the remaining network nodes and network topology to predict the possibility of connection between the remaining two nodes that have not yet generated connection edges;

[0034] Combined with the growth of enterprises in different industries, according to the distribution law of enterprise registered capital, registered place, and enterprise type information, seed enterprises are randomly generated through the Monte Carlo method, and combined with the temporal characteristics of the evolution of regional industrial networks, the connection relationship of newly added nodes is predicted;

[0035] Based on the enterprise relationship network and the connection relationship of newly added nodes, a regional industrial network evolution trend forecast map is drawn.

[0036] Compared with the prior art, the analysis method for regional industrial network evolution of the present invention has the following beneficial effects:

[0037] The present invention uses time series similarity to improve the industrial development trend prediction model, which can effectively reduce the error of short-term prediction and ensure the prediction of long-term growth trend; it uses enterprise-related data of multiple dimensions, including basic information, business activities, and external environment characteristics of the enterprise, with rich data dimensions and more scientific analysis results. By integrating similarity indicators of three dimensions, such as node feature similarity, distance proximity, and relationship proximity, the accuracy and robustness of link prediction can be effectively improved. The basic data used in the present invention mainly come from the basic data of the Market Supervision Administration and the open source data of the Internet. No data from special sources is used. The use conditions of the method are more general and easy to promote and use in different cities. Using the Monte Carlo method, the evolution law of the industrial network can be dynamically simulated by randomly producing new seed enterprises.

[0038] In a second aspect, an embodiment of the present invention provides an analysis system for regional industrial network evolution, including:

[0039] The enterprise information database construction module is used to obtain the enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for the analysis of the evolution of the industrial network in a specific region;

[0040] The enterprise growth prediction module is used to extract the enterprise growth time series data of different industries in the region in turn based on the enterprise information database and the industry classification of 96 major categories of the national economy, and to build a regional industry-specific growth trend prediction model to predict the future growth of enterprises in different industries in a specific region;

[0041] The enterprise extinction risk prediction module is used to construct an enterprise feature vector and a positive and negative sample data set from the two dimensions of enterprise business activity characteristics and external environment characteristics based on the enterprise information database; divide the positive and negative sample data sets into training sets and test sets; use the training set and the test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the enterprise extinction risk;

[0042] The industrial network evolution prediction module is used to build an enterprise relationship network based on the enterprise information database, remove enterprises with high extinction risks, predict the potential relationships between the remaining nodes through link prediction methods based on the remaining industrial network nodes and relationships, and realize industrial network evolution prediction in combination with the growth of enterprises in different industries.

[0043] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a processor, a communication interface, a memory and a bus, wherein the processor, the communication interface and the memory communicate with each other through the bus, and the processor can call logic instructions in the memory to execute the steps of the method provided in the first aspect.

[0044] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the analysis method for regional industrial network evolution as described in the first aspect.

[0045] Compared with the prior art, the beneficial effects of the analysis system, electronic device and storage medium for regional industrial network evolution of the present invention are the same as the analysis method for regional industrial network evolution described in the first aspect, so they will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 is a flow chart of an analysis method for regional industrial network evolution in an embodiment of the present invention;

[0048] Figure 2 is a structural block diagram of an analysis system for regional industrial network evolution in an embodiment of the present invention;

[0049] Figure 3 1 is a structural block diagram of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0051] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with general skills in the technical field to which this application belongs. The words "one", "a", "the", "these" and the like in this application do not indicate a quantitative limitation, and they may be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" may mean: A exists alone, A and B exist at the same time, and B exists alone. Generally, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0052] In an embodiment of the present invention, a method for analyzing the evolution of a regional industrial network is provided. Figure 1 is a flow chart of the analysis method for regional industrial network evolution of the present invention, such as Figure 1 As shown, the process includes the following steps:

[0053] S100, obtaining enterprise registration information database, enterprise business activity database and enterprise external environment data and constructing an enterprise information database for specific regional industrial network evolution analysis;

[0054] It should be noted that, first of all, based on the company's registration data, annual report data, change data, administrative penalty data, recruitment data, bidding data, shareholder data, product data, patent data and other data, the company data is integrated according to the company's unified credit code and company name to form an enterprise information database; combined with Internet open source data, POI (point of interest) data, transportation infrastructure data, office leasing, population density data, company litigation data and other external data in the target area are collected to form the company's external environment database, and finally a basic database for regional industrial analysis is formed.

[0055] Specifically, in this embodiment, obtaining the enterprise registration information database, the enterprise business activity database and the enterprise external environment data and constructing an enterprise information database for the analysis of the evolution of the industrial network in a specific region includes:

[0056] Based on the enterprise registration data, the unified credit code is used as the unique identifier to clean and pre-process the registration data to form an enterprise registration information database;

[0057] According to the unified credit code of the enterprise, the enterprise data are aligned and integrated to obtain the enterprise business activity database; it should be noted that the enterprise data includes the annual report information, change information, abnormal operation information, administrative penalty information, product information, intellectual property declaration information, bidding information, recruitment information, qualification information and administrative license information and other enterprise data.

[0058] Obtain open source data from the Internet, including regional POI data, population density data, transportation facility data, office leasing data, statistical yearbook data, and corporate litigation data, and obtain corporate external environment data through geocoding based on the address information contained in the open source data;

[0059] It should be noted that open source data on the Internet can be obtained through Internet crawler technology.

[0060] Based on the enterprise registration information database, enterprise business activity database and enterprise external environment data, an enterprise information database for specific regional industrial network evolution analysis is constructed.

[0061] S200, based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, sequentially extract the enterprise growth time series data of different industries in the region, and build a regional industry-specific growth trend prediction model to predict the future growth of enterprises in different industries in a specific region;

[0062] Based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional sub-industry growth trend prediction model is constructed to predict the future regional industrial growth trend. Then, based on the existing industrial structure of the region, the industrial growth amount of the sub-industry of the region is allocated, and the distribution map of the regional industrial growth amount is output; the present invention uses time series similarity to improve the industrial development trend prediction model, which can effectively reduce the error of short-term prediction and ensure the prediction of long-term growth trend.

[0063] Specifically, based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-specific growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region, including:

[0064] Based on the enterprise information database, combined with the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn. ; k represents the industry type.

[0065] Based on the enterprise growth time series data, a regional and industry growth trend forecasting model is constructed; as shown below:

[0066] ;

[0067] The numerical solution of the classical growth model is obtained by the differential equation numerical method, and the initial analysis results are obtained. ;

[0068] ;

[0069] Combining initial analysis results Establish The timing analysis model is as follows:

[0070] ;

[0071] Among them, the parameters α 0. α 1. α 2 and α 3 is the parameter to be estimated.

[0072] The parameters of the time series analysis model are obtained by the time series parameter estimation method ( α 0. α 1. α 2 and α 3);

[0073] Use the time series analysis model to predict the future growth of enterprises in different industries in a specific region.

[0074] This enables prediction of the growth of enterprises in different industries in a specific region in the next 1, 2, …, N years.

[0075] S300, based on the enterprise information database, constructing enterprise feature vectors and constructing positive and negative sample data sets from two dimensions: enterprise business activity characteristics and external environment characteristics; dividing the positive and negative sample data sets into training sets and test sets; using the training set and the test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the risk of enterprise extinction;

[0076] In this embodiment, based on the enterprise information database, the enterprise feature vector is constructed from two dimensions: the enterprise business activity characteristics and the external environment characteristics, forming an enterprise feature vector containing 42 features. According to the business status field in the enterprise information database, the enterprises that are cancelled or revoked and the enterprises that are opened or in existence are proposed from the enterprise registration information to form a positive and negative sample data set. Then, two machine learning algorithms are selected to perform model training respectively, and a voting model based on model fusion is constructed to predict the risk of enterprise extinction. Multiple dimensions of enterprise-related data are used, including the basic information, business activities, and external environment characteristics of the enterprise. The data dimensions are rich and the analysis results are more scientific.

[0077] Specifically, based on the enterprise information database, an enterprise feature vector is constructed from two dimensions: enterprise business activity characteristics and external environment characteristics, and a positive and negative sample data set is constructed; the positive and negative sample data sets are divided into a training set and a test set; the XGBOOST model and the LightGBM model are trained using the training set and the test set, respectively, and the two trained models are used to predict the risk of enterprise extinction, including:

[0078] Based on the enterprise information database, the characteristics of the enterprise's business activities and external environment are extracted. The characteristics of the enterprise's business activities include static characteristics and dynamic characteristics; static characteristics include the name, type, establishment time, registered capital, registered place, shareholders and main managers of the enterprise, etc. Dynamic characteristics include main business income, profit margin, number of employees, number of bidding, number of intellectual property applications, number of recruitment, number of changes, number of administrative penalties, number of abnormal operations, etc., forming a feature vector containing 28 characteristics reflecting the characteristics of the enterprise's business situation. According to the address information of the enterprise, through geographic space matching, various types of urban land supply, average salary, number of recruitment positions, population density, traffic accessibility, population density, urban vitality, number of patents, number of cases involved in litigation and other external environmental characteristics of the area where the enterprise is located are extracted, forming a feature vector containing 14 characteristics reflecting the external environment characteristics of the enterprise.

[0079] Construct positive and negative sample sets based on the business status of the enterprise, and divide the enterprise into training set and test set by random sampling;

[0080] Specifically, based on the enterprise information database and combined with the enterprise business status field, enterprises with a status of revoked or cancelled are extracted as defunct enterprises; enterprises with a status of opened or continuing are extracted as normal enterprises to form positive and negative sample sets, and the enterprises are divided into training sets and test sets according to the 80 / 20 principle through random sampling.

[0081] The XGBOOST model and LightGBM model are trained using the training set and test set respectively;

[0082] Based on the trained XGBOOST model and LightGBM model, the enterprise extinction risk prediction is realized. The enterprise extinction risk prediction results include high extinction risk and low extinction risk.

[0083] Based on the trained XGBOOST model and LightGBM model, the risk of enterprise extinction is predicted. The calculation method is as follows:

[0084] ;

[0085] Among them, Pi represents the extinction probability of enterprise i, which is the average of the predicted probabilities of the XGBOOST model and the LightGBM model.

[0086] Among them, Pi greater than or equal to 0.5 is considered a high extinction risk, and Pi less than 0.5 is considered a low extinction risk.

[0087] S400, constructing an enterprise relationship network based on the enterprise information database, removing enterprises with high extinction risk, predicting the potential relationship between the remaining nodes through a link prediction method based on the nodes and relationships of the remaining industrial network, and realizing the industrial network evolution prediction in combination with the growth of enterprises in different industries. By integrating the similarity index of three dimensions, namely, node feature similarity, distance proximity, and relationship proximity, the accuracy and robustness of link prediction can be effectively improved.

[0088] Based on the enterprise information database, the investment relationship, supply chain relationship, industrial chain relationship, cooperation relationship and other relationship types of enterprises are extracted in turn to form an enterprise relationship network. Then, the enterprises with high extinction risk are eliminated, and based on the nodes and relationships of the remaining industrial network, the potential relationship between the remaining nodes is predicted through the link prediction method. The number of new enterprises in each industry is predicted, and the connection relationship between the new nodes and the existing nodes is predicted, and finally the evolution prediction of the industrial network is realized.

[0089] Specifically, an enterprise relationship network is constructed based on the enterprise information database, enterprises with high extinction risk are removed, and based on the nodes and relationships of the remaining industrial network, the potential relationships between the remaining nodes are predicted by the link prediction method, and the industrial network evolution prediction is realized in combination with the growth of enterprises in different industries, including:

[0090] Extract enterprise relationships in sequence, and use the entropy method to calculate the relationship weights of enterprises to form an enterprise relationship network. Enterprise relationships include investment relationships, supply chain relationships, industrial chain relationships, and cooperative relationships.

[0091] ;

[0092] in, Indicates the enterprise i,j Relationship k The weight of Indicates the enterprise i,j , in the relationship k The proportion of the following.

[0093] Eliminate enterprises with high extinction risk and their connection edges, and use the link prediction method based on the remaining network nodes and network topology to predict the possibility of connection between the remaining two nodes that have not yet generated connection edges; the calculation formula is as follows:

[0094] ;

[0095] in, ω 1 , ω 2 , ω 3 Represents the parameter to be estimated. Indicates the enterprise i,j The probability of a connection, Indicates the enterprise i,j The cosine similarity between Indicates the enterprise i,j The distance between and Respectively represent enterprises i,j The number of first-order neighbors and second-order neighbors between them.

[0096] Combined with the growth of enterprises in different industries, according to the distribution law of enterprise registered capital, registered place, and enterprise type information, seed enterprises are randomly generated through the Monte Carlo method, and combined with the temporal characteristics of the evolution of the regional industrial network, the connection relationship of the newly added nodes is predicted; the calculation formula is as follows:

[0097] ;

[0098] Among them, ξ 1 , ξ 2 , ξ 3 Represents the parameter to be estimated. Indicates a newly added node j With existing nodes i The probability of connection, Representation Nodei The node degree, Indicates the enterprise i,j The cosine similarity between It indicates the proportion of new connections generated by newly added nodes in the regional industrial network in the past five years. The Monte Carlo method can be used to dynamically simulate the evolution of the industrial network by randomly generating new seed enterprises.

[0099] Based on the enterprise relationship network and the connection relationship of newly added nodes, a regional industrial network evolution trend forecast map is drawn.

[0100] It should be noted that the basic data used in the present invention mainly comes from the basic data of the Market Supervision Administration and the open source data on the Internet. No data from special sources are used. The conditions for using the method are more general and can be easily promoted and used in different cities.

[0101] The embodiment of the present invention also provides an analysis system for the evolution of regional industrial networks, which is used to implement the above method embodiments, and will not be repeated hereafter. The terms "module", "unit", "subunit", etc. used below can implement a combination of software and / or hardware for predetermined functions. Although the systems described in the following embodiments are preferably implemented in software, the implementation of hardware or a combination of software and hardware is also possible and conceivable.

[0102] like Figure 2 As shown, Figure 2 It is a structural block diagram of the analysis system for regional industrial network evolution in the present invention, and the system includes:

[0103] The enterprise information database construction module 101 is used to obtain the enterprise registration information database, the enterprise business activity database and the enterprise external environment data and construct an enterprise information database for the analysis of the evolution of the industrial network in a specific region;

[0104] The enterprise growth prediction module 102 is used to extract the enterprise growth time series data of different industries in the region in turn based on the enterprise information database and the industry classification of 96 major categories of the national economy, and to build a regional industry-specific growth trend prediction model to predict the future growth of enterprises in different industries in a specific region;

[0105] The enterprise extinction risk prediction module 103 is used to construct an enterprise feature vector and a positive and negative sample data set from two dimensions, namely, the enterprise business activity characteristics and the external environment characteristics, based on the enterprise information database; divide the positive and negative sample data sets into a training set and a test set; use the training set and the test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the enterprise extinction risk;

[0106] The industrial network evolution prediction module 104 is used to construct an enterprise relationship network based on the enterprise information database, remove enterprises with high extinction risks, and predict the potential relationships between the remaining nodes based on the nodes and relationships of the remaining industrial network through a link prediction method, and realize the industrial network evolution prediction in combination with the growth of enterprises in different industries.

[0107] Figure 3 A structural block diagram of an electronic device provided by an embodiment of the present invention, such as Figure 3 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the following method:

[0108] Obtain enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for specific regional industrial network evolution analysis;

[0109] Based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-specific growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region;

[0110] Based on the enterprise information database, construct enterprise feature vectors and construct positive and negative sample data sets from two dimensions: enterprise business activity characteristics and external environment characteristics; divide the positive and negative sample data sets into training sets and test sets; use the training set and test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the risk of enterprise extinction;

[0111] Based on the enterprise information database, an enterprise relationship network is constructed to remove enterprises with high extinction risks. Based on the nodes and relationships of the remaining industrial network, the potential relationships between the remaining nodes are predicted through link prediction methods, and the industrial network evolution prediction is realized in combination with the growth of enterprises in different industries.

[0112] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0113] An embodiment of the present invention further provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method provided in the above embodiment is implemented.

[0114] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An analysis method for regional industrial network evolution, characterized in that: The analysis method comprises: Obtain enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for specific regional industrial network evolution analysis; Based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-specific growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region; Based on the enterprise information database, construct enterprise feature vectors and construct positive and negative sample data sets from two dimensions: enterprise business activity characteristics and external environment characteristics; divide the positive and negative sample data sets into training sets and test sets; use the training set and test set to train the XGBOOST model and the LightGBM model respectively, and the two trained models are used to predict the risk of enterprise extinction; Based on the enterprise information database, an enterprise relationship network is constructed, and enterprises with high extinction risk are removed. Based on the nodes and relationships of the remaining industrial network, the potential relationships between the remaining nodes are predicted by the link prediction method, and the industrial network evolution prediction is realized by combining the growth of enterprises in different industries, including: Extract enterprise relationships in sequence, and use the entropy method to calculate the relationship weights of enterprises to form an enterprise relationship network. Enterprise relationships include investment relationships, supply chain relationships, industrial chain relationships, and cooperative relationships. Eliminate enterprises with high extinction risk and their connection edges, and use link prediction methods based on the remaining network nodes and network topology to predict the possibility of connection between the remaining two nodes that have not yet generated connection edges; Combined with the growth of enterprises in different industries, according to the distribution law of enterprise registered capital, registered place, and enterprise type information, seed enterprises are randomly generated through the Monte Carlo method, and combined with the temporal characteristics of the evolution of regional industrial networks, the connection relationship of newly added nodes is predicted; Based on the enterprise relationship network and the connection relationship of newly added nodes, a regional industrial network evolution trend forecast map is drawn.

2. According to claim 1, the analysis method for regional industrial network evolution is characterized by: Obtaining the enterprise registration information database, enterprise business activity database and enterprise external environment data and constructing an enterprise information database for specific regional industrial network evolution analysis, including: Based on the enterprise registration data, the unified credit code is used as the unique identifier to clean and pre-process the registration data to form an enterprise registration information database; According to the unified credit code of the enterprise, the enterprise data is aligned and integrated to obtain the enterprise business activity database; Obtain open source data from the Internet, including regional POI data, population density data, transportation facility data, office leasing data, statistical yearbook data, and corporate litigation data, and obtain corporate external environment data through geocoding based on the address information contained in the open source data; Based on the enterprise registration information database, enterprise business activity database and enterprise external environment data, an enterprise information database for specific regional industrial network evolution analysis is constructed.

3. The method for analyzing the evolution of regional industrial networks according to claim 1 is characterized in that: Based on the enterprise information database, according to the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn, and a regional industry-specific growth trend prediction model is constructed to predict the future growth of enterprises in different industries in a specific region, including: Based on the enterprise information database, combined with the industry classification of 96 major categories of the national economy, the time series data of enterprise growth in different industries in the region are extracted in turn. ; Construct a regional and industry-specific growth trend forecasting model based on enterprise growth time series data; The numerical solution of the classical growth model is obtained by the differential equation numerical method, and the initial analysis results are obtained. ; Combining initial analysis results Establish Timing analysis model of The parameters of the time series analysis model are obtained through the time series parameter estimation method; Use the time series analysis model to predict the future growth of enterprises in different industries in a specific region.

4. According to claim 1, the method for analyzing the evolution of regional industrial networks is characterized in that: Based on the enterprise information database, the enterprise feature vector is constructed from the two dimensions of enterprise business activity characteristics and external environment characteristics, and the positive and negative sample data sets are constructed; the positive and negative sample data sets are divided into training sets and test sets; the XGBOOST model and the LightGBM model are trained using the training set and the test set respectively, and the two trained models are used to predict the risk of enterprise extinction, including: Based on the enterprise information database, extracting enterprise business activity characteristics and external environment characteristics, wherein the enterprise business activity characteristics include static characteristics and dynamic characteristics; Construct positive and negative sample sets based on the business status of the enterprise, and divide the enterprise into training set and test set by random sampling; The XGBOOST model and LightGBM model are trained using the training set and test set respectively; Based on the trained XGBOOST model and LightGBM model, the enterprise extinction risk prediction is realized. The enterprise extinction risk prediction results include high extinction risk and low extinction risk.

5. The method for analyzing the evolution of regional industrial networks according to claim 4 is characterized in that: Based on the trained XGBOOST model and LightGBM model, the risk of enterprise extinction is predicted. The calculation method is as follows: ； Among them, Pi represents the extinction probability of enterprise i, which is the average of the predicted probabilities of the XGBOOST model and the LightGBM model; Among them, Pi greater than or equal to 0.5 is considered a high extinction risk, and Pi less than 0.5 is considered a low extinction risk.

6. An analysis system for regional industrial network evolution, characterized in that: include: The enterprise information database construction module is used to obtain the enterprise registration information database, enterprise business activity database and enterprise external environment data and build an enterprise information database for the analysis of the evolution of the industrial network in a specific region; The enterprise growth prediction module is used to extract the enterprise growth time series data of different industries in the region in turn based on the enterprise information database and the industry classification of 96 major categories of the national economy, and to build a regional industry-specific growth trend prediction model to predict the future growth of enterprises in different industries in a specific region; An enterprise extinction risk prediction module is used to construct an enterprise feature vector and a positive and negative sample data set based on the enterprise information database from two dimensions: enterprise business activity characteristics and external environment characteristics; The positive and negative sample data sets are divided into training sets and test sets. The XGBOOST model and the LightGBM model are trained using the training set and the test set respectively. The two trained models are used to predict the risk of enterprise extinction. The industrial network evolution prediction module is used to construct an enterprise relationship network based on the enterprise information database, remove enterprises with high extinction risks, and predict the potential relationship between the remaining nodes through the link prediction method based on the remaining nodes and relationships of the industrial network, and realize the industrial network evolution prediction in combination with the growth of enterprises in different industries, including: Extract enterprise relationships in sequence, and use the entropy method to calculate the relationship weights of enterprises to form an enterprise relationship network. Enterprise relationships include investment relationships, supply chain relationships, industrial chain relationships, and cooperative relationships. Eliminate enterprises with high extinction risk and their connection edges, and use link prediction methods based on the remaining network nodes and network topology to predict the possibility of connection between the remaining two nodes that have not yet generated connection edges; Combined with the growth of enterprises in different industries, according to the distribution law of enterprise registered capital, registered place, and enterprise type information, seed enterprises are randomly generated through the Monte Carlo method, and combined with the temporal characteristics of the evolution of regional industrial networks, the connection relationship of newly added nodes is predicted; Based on the enterprise relationship network and the connection relationship of newly added nodes, a regional industrial network evolution trend forecast map is drawn.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the analysis method for regional industrial network evolution according to any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the analysis method for regional industrial network evolution as described in any one of claims 1 to 5 are implemented.

Images