Enterprise industry classification method and system based on associated enterprise information and BERT model

Abstract

The invention discloses an enterprise industry classification method and system based on associated enterprise information and a BERT model, and relates to the technical field of computers. The method comprises the following steps: acquiring enterprise information of a to-be-classified enterprise; enterprise keywords of the to-be-classified enterprises are extracted according to the enterprise information, and an enterprise keyword set of the to-be-classified enterprises is determined based on the enterprise keywords; respectively calculating the similarity between the enterprise keyword set and each preset industry keyword set through a pre-trained BERT model; the industry categories corresponding to the industry keyword sets with the similarity larger than a similarity threshold value are selected as alternative industries to which the industry categories belong; if the alternative industry is not unique, determining an associated enterprise of the to-be-classified enterprise, processing the associated enterprise information of the associated enterprise and the similarity of each alternative industry according to a preset comprehensive industry score calculation rule, and calculating to obtain a comprehensive industry score of each alternative industry; and the industry with the highest score is selected as the belonging industry, so that the accuracy of enterprise industry classification is improved.

Description

[0001] This application is a divisional application of patent application No. 202310869145.4, filed on July 14, 2023, entitled "A method, system, device and medium for classifying enterprise industries based on big data". Technical Field

[0002] This application relates to the field of computer technology, and in particular to a method and system for classifying enterprise industries based on related enterprise information and the BERT model. Background Technology

[0003] The National Economic Industry Classification is a standardized method for classifying various industries according to their different characteristics of production and operation activities. Currently, the National Economic Industry Classification issued by the National Bureau of Statistics is the most commonly used industry classification standard.

[0004] For every company, its industry tag is a crucial field, effectively reflecting its main business operations. Therefore, a company database needs to categorize companies by industry to determine their industry tags.

[0005] Current methods for classifying companies by industry typically rely on a single or limited set of indicators. This approach is susceptible to limitations imposed by the companies themselves, leading to inaccurate industry classifications. Summary of the Invention

[0006] To improve the accuracy of enterprise industry classification, this application provides an enterprise industry classification method and system based on related enterprise information and the BERT model.

[0007] Firstly, this application provides a method for classifying enterprise industries based on big data, the method comprising the following steps: Obtain enterprise information for companies to be classified; Extract enterprise keywords from the enterprise information to be classified, and determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; The similarity between the enterprise keyword set and the pre-set industry keyword sets is calculated using a pre-set similarity calculation model. The industry categories corresponding to the set of industry keywords with similarity greater than the similarity threshold are selected as the candidate industries to which the enterprise to be classified belongs.

[0008] By employing the above technical solution, a set of enterprise keywords is determined based on the enterprise information of the enterprise to be classified, and a set of industry keywords is determined based on the industry keywords of each industry category. The similarity between the enterprise keyword set and the industry keyword set is used to determine the candidate industries to which the enterprise to be classified belongs. In determining the enterprise keyword set and industry keyword set, multi-dimensional enterprise information and industry data are considered, thus providing a more accurate description of the enterprise to be classified and each industry category, which helps improve the accuracy of enterprise industry classification.

[0009] Optionally, after selecting the industry categories corresponding to the industry keyword set with similarity greater than a similarity threshold as the candidate industries to which the enterprise to be classified belongs, the method further includes: Determine whether the candidate industry to which the enterprise to be classified belongs is unique; If not, then identify the related companies of the company to be classified, and obtain the related company information of each related company. The related company information includes related company relationship information and related company industry classification information. Based on the associated enterprise information, determine the industry to which the enterprise to be classified belongs from among several candidate industries.

[0010] By adopting the above technical solution, when classifying enterprises into different industries, the industry affiliation may be ambiguous due to the inherent attributes of the enterprise itself. In such cases, multiple alternative industries may be considered, with the similarity of these alternative industries exceeding a similarity threshold. Therefore, based on the information of the related enterprises of the enterprise to be classified, the enterprise can be further classified, thereby determining its accurate industry affiliation and improving the accuracy of industry classification.

[0011] Optionally, after determining whether the candidate industry of the enterprise to be classified is unique, the method further includes: If so, the candidate industry will be taken as the industry of the enterprise to be classified.

[0012] By adopting the above technical solution, when there is only one candidate industry for the enterprise to be classified, it indicates that the enterprise to be classified has distinct industry characteristics. In this case, the candidate industry is directly used as the industry of the enterprise to be classified, thus completing the industry classification of the enterprise to be classified.

[0013] Optionally, before obtaining the enterprise information of the enterprises to be classified, an industry keyword set creation method is also included, wherein the industry keyword set creation method specifically includes: Obtain the National Economic Industry Classification document; Create multiple sets of industry keywords corresponding to the industry categories specified in the National Economic Industry Classification document; Obtain the industry keywords corresponding to the industry categories of each industry keyword set; The industry keywords for each industry category are stored in their respective industry keyword sets, thus completing the creation of each industry keyword set.

[0014] By adopting the above technical solution, an industry keyword set is created according to the industry categories specified in the National Economic Industry Classification document, thereby describing each industry category through the industry keyword set.

[0015] Optionally, the acquisition of industry keywords for the industry categories corresponding to each of the aforementioned industry keyword sets specifically includes: For a given set of industry keywords, the first industry keyword is obtained based on the industry annotations of the industry categories corresponding to the set of industry keywords in the National Economic Industry Classification document. Obtain the enterprise information of companies in the same industry from a pre-set enterprise database, wherein the industry to which the companies in the same industry belong is the same as the industry category corresponding to the industry keyword set; Obtain second industry keywords based on the company information of the companies in the same industry.

[0016] By adopting the above technical solution, the National Economic Industry Classification document, a nationally mandated standard for enterprise industry classification, provides industry annotations for the industry categories corresponding to the industry keyword set. These annotations explain the characteristics of the industry category and offer a relatively good description of it. The first industry keyword obtained from the industry annotations can provide a global explanation of the industry category, thus achieving an abstract description. However, the industry annotations for the industry categories corresponding to the industry keyword set in the National Economic Industry Classification document do not provide a comprehensive description of the industry category. Extracting the second industry keyword from enterprise information within the same industry allows for a further description of the industry category, improving the interpretability of the industry keyword set for the corresponding industry category.

[0017] Optionally, determining the industry to which the enterprise to be classified belongs from a plurality of candidate industries based on the associated enterprise information specifically includes: The similarity between the related enterprise information and each of the candidate industries is processed according to the preset comprehensive industry scoring calculation rules to calculate the comprehensive industry score of each of the candidate industries. The candidate industry with the highest comprehensive industry score is selected as the industry to which the enterprise to be classified belongs.

[0018] By adopting the above technical solution, each candidate industry is evaluated based on a comprehensive industry score. The comprehensive industry score further reflects the similarity between each candidate industry and the enterprise to be classified, thereby selecting the industry to which the enterprise to be classified belongs from multiple candidate industries and further improving the accuracy of enterprise industry classification.

[0019] Optionally, in processing the similarity between the related enterprise information and each of the candidate industries according to the preset comprehensive industry scoring calculation rules, and calculating the comprehensive industry score of each candidate industry, the specific steps include: For a candidate industry, the similarity between the candidate industry and the enterprise to be classified is used as the first industry score; Among the related enterprises of the enterprise to be classified, identify the related enterprises in the same industry category as the candidate enterprise; Calculate the second industry score based on the related enterprise information of the related enterprises in the same industry; The first industry score and the second industry score are weighted and calculated to complete the calculation of the comprehensive industry score of the candidate industry.

[0020] By adopting the above technical solution, the comprehensive industry score considers both the similarity between the candidate industry and the enterprise to be classified, and related enterprises in the same industry for each candidate industry. These related enterprises belong to the candidate industries for which the comprehensive industry score needs to be calculated, and are also related to the enterprise to be classified. By using related enterprises in the same industry, the possibility of which candidate industry the enterprise to be classified belongs to can be further described, thus making the calculated comprehensive industry score of the candidate industry more convincing.

[0021] A second aspect of this application provides an enterprise industry classification system based on big data, the system comprising the following modules: The enterprise information acquisition module is used to acquire enterprise information of enterprises to be classified. The enterprise keyword extraction module is used to extract enterprise keywords of the enterprise to be classified based on the enterprise information, and to determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; The similarity calculation module is used to calculate the similarity between the enterprise keyword set and the preset industry keyword sets respectively using a preset similarity calculation model; The alternative industry determination module is used to select the industry categories corresponding to the set of industry keywords with similarity greater than a similarity threshold as alternative industries for the enterprise to be classified.

[0022] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of the first aspects.

[0023] A fourth aspect of this application provides a computer-readable storage medium storing instructions that, when executed, perform the steps of the method as described in any one of the first aspects. In summary, one or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Based on the enterprise information of the companies to be classified, determine the enterprise keyword set; based on the industry keywords of each industry category, determine the industry keyword set for each industry category; and by calculating the similarity between the enterprise keyword set and the industry keyword set, determine the candidate industries to which the companies to be classified belong. Considering multi-dimensional enterprise information and industry data allows for a more accurate description of the companies to be classified and their respective industry categories, thus improving the accuracy of enterprise industry classification.

[0024] 2. When determining the industry keyword set, we should consider both the industry annotations of the industry category corresponding to the industry keyword set in the National Economic Industry Classification document and the enterprise information of companies in the same industry, so that the industry keyword set can provide a clearer, more comprehensive and accurate description of the industry category.

[0025] 3. When the industry attributes of the enterprise to be classified are relatively ambiguous, consider the related enterprises of the enterprise to be classified, and further classify the enterprise to be classified by industry based on the information of the related enterprises, so as to further improve the accuracy of the industry classification. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a big data-based enterprise industry classification method provided in an embodiment of this application.

[0027] Figure 2 This is a schematic diagram illustrating the generation of the first industry score and the second industry score in a big data-based enterprise industry classification method provided in this application embodiment.

[0028] Figure 3 This is a schematic diagram of the structure of an enterprise industry classification system based on big data disclosed in an embodiment of this application.

[0029] Figure 4 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0030] Explanation of reference numerals in the attached diagram: 301, Enterprise information acquisition module; 302, Enterprise keyword extraction module; 303, Similarity calculation module; 304, Candidate industry determination module; 400, Electronic device; 401, Processor; 402, Communication bus; 403, User interface; 404, Network interface; 405, Memory. Detailed Implementation

[0031] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0032] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0033] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0034] Reference Figure 1 This application provides a big data-based enterprise industry classification method, which specifically includes the following steps: S10: Obtain enterprise information of the enterprise to be classified; Specifically, the enterprise information of the enterprises to be classified is obtained from the enterprise database. Specifically, the enterprise information includes enterprise name information, main business information, enterprise profile information, and other business registration information related to the enterprises to be classified.

[0035] It should be noted that the company information to be classified was entered into the company database before the industry classification was performed. This database contains a large amount of company data. The company database includes not only companies to be classified, but also some companies that have already been classified by industry and have industry tags generated based on the classification results.

[0036] S20: Extract enterprise keywords from the enterprise information to be classified, and determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; Specifically, for the original text of the obtained enterprise information, NLP technology is used to extract enterprise keywords for the enterprises to be classified. There are multiple extracted enterprise keywords. After the extraction of enterprise keywords is completed, the set of enterprise keywords for the enterprises to be classified is determined based on the enterprise keywords.

[0037] Specifically, extracting enterprise keywords from enterprise information based on NLP technology first requires preprocessing the original text of the enterprise information. This preprocessing includes word segmentation, stop word removal, and part-of-speech tagging. After obtaining the preprocessed text, enterprise keywords can be extracted using various keyword extraction algorithms, including but not limited to TextRank, LDA, RAKE, and TF-IDF. In one feasible embodiment of this application, enterprise keyword extraction can be performed using natural language processing toolkits such as NLTK, spaCy, or gensim. In another feasible embodiment, enterprise keyword extraction can also be based on a deep learning keyword extraction model.

[0038] Enterprise keywords can describe the enterprises to be classified and reflect the main characteristics of their business operations. After extracting the enterprise keywords, an enterprise keyword set for the enterprises to be classified is established.

[0039] S30: Calculate the similarity between the enterprise keyword set and the pre-set keyword sets of each industry using a pre-set similarity calculation model; Specifically, after determining the enterprise keyword set of the company to be classified and the industry keyword set of each industry category, the similarity between the enterprise keyword set and each industry keyword set is calculated using a pre-set similarity calculation model.

[0040] The industry keyword sets were created before the enterprise industry classification. These sets describe various industry categories, and their creation process involves: obtaining the National Economic Industry Classification document; creating multiple industry keyword sets corresponding to the industry categories specified in the National Economic Industry Classification document; obtaining the industry keywords for each industry category corresponding to each industry keyword set; and storing the industry keywords for each industry category into their respective industry keyword sets, thus completing the creation of each industry keyword set.

[0041] The industry categories are specifically set according to the "National Economic Industry Classification and Codes" (GB / T 4754-2017) (hereinafter referred to as the National Economic Industry Classification Document) issued by the State Council on June 30, 2017. The National Economic Industry Classification Document stipulates the classification and codes for the entire society, applicable to the classification of economic activities in national statistical, planning, fiscal, tax, and industrial and commercial systems, and used for information processing. The National Economic Industry Classification Document divides the entire society's industries into four levels: categories, major categories, intermediate categories, and minor categories, including 20 categories, 97 major categories, 473 intermediate categories, and 1380 minor categories.

[0042] In one feasible embodiment of this application, the number and types of industry categories are set to the number and types of subcategories in the National Economic Industry Classification document, that is, 473 industry categories are set; in other embodiments of this application, the industry categories can also be set based on the knowledge of those skilled in the art or actual business needs.

[0043] Obtain industry keywords for each industry category, and determine the industry keyword set for each industry category based on these keywords. Each industry keyword set includes a first industry keyword and a second industry keyword, both of which are used to describe the industry category corresponding to their respective keyword set.

[0044] For a set of industry keywords for an industry category, the first industry keyword is determined based on the industry annotation of the industry category corresponding to the set of industry keywords in the National Economic Industry Classification document. The industry annotation of the industry category in the National Economic Industry Classification document is obtained, and the first industry keyword is extracted based on NLP technology.

[0045] For a set of industry keywords for an industry category, the second industry keywords included are determined based on the enterprise information of companies within the same industry category. Companies within the same industry category refer to those belonging to the same industry. As described above, there are categorized companies in the enterprise database, and the industries of these categorized companies have been determined. The second industry keywords for this industry category are extracted based on the enterprise information of the companies within the categorized companies. Similarly, the extraction method for the second industry keywords is also based on NLP technology. In one feasible embodiment of this application, since there are many companies within the same industry, extracting the second industry keywords from all companies within an industry category would result in an excessively large data volume. Therefore, based on the expert opinions of those skilled in the art, several of the most representative companies are selected from among the many companies within the same industry as the source for extracting the second industry keywords.

[0046] For a set of enterprise keywords and a set of industry keywords, the two sets are first encoded using a pre-defined similarity calculation model, resulting in enterprise keyword vectors for the enterprise set and industry keyword vectors for the industry set. Then, the cosine similarity between the enterprise keyword vectors and the industry keyword vectors is calculated, and the result is used as the similarity between the enterprise keyword set and the industry keyword set. In one feasible embodiment of this application, the similarity calculation model can be a BERT model, and the similarity calculation model has been pre-trained before performing the similarity calculation.

[0047] S40: Select the industry categories corresponding to the set of industry keywords with similarity greater than the similarity threshold as the candidate industries to be classified for the enterprise; Specifically, after calculating the similarity between the enterprise keyword set of the enterprise to be classified and the industry keyword set of each industry category, the obtained similarity is compared with the preset similarity threshold. The industry category corresponding to the industry keyword set with a similarity greater than the similarity threshold is selected as the candidate industry to which the enterprise to be classified belongs.

[0048] S50: Determine whether the alternative industry to which the enterprise to be classified belongs is unique; S60: If there is only one alternative industry for the enterprise to be classified, then the alternative industry shall be used as the industry for the enterprise to be classified. Specifically, if there is only one alternative industry to which the enterprise to be classified belongs, it means that the characteristics of the enterprise to be classified are significantly similar to those of the alternative industry, and the enterprise to be classified belongs to the alternative industry. In this case, the unique alternative industry will be taken as the industry to which the enterprise to be classified belongs.

[0049] S61: If the candidate industries of the enterprise to be classified are not unique, then determine the related enterprises of the enterprise to be classified and obtain the related enterprise information of each related enterprise. Specifically, since the set similarity threshold is a fixed value, there may be situations where the similarity between the keyword set of the company to be classified and multiple industry keyword sets is greater than the set similarity threshold. If the similarity between the keyword set of the company and multiple industry keyword sets is greater than the set similarity threshold, it means that the business type of the company to be classified is relatively complex and has a certain degree of similarity with multiple candidate industries. In this case, it is difficult to classify the company into an industry solely based on the keyword set of the company to be classified.

[0050] When there is multiple candidate industries for a company to be classified, the company's related companies are searched in the enterprise database. Related companies refer to enterprises that have cooperative, competitive, or supply relationships with the company to be classified, and there may be multiple related companies. After identifying the related companies of the company to be classified, the related company information for each related company is obtained. The related company information includes related company relationship information, related company industry classification information, and information about other related companies. Among them, the related company relationship information describes the relationship between the related company and the company to be classified, and the related company industry classification information describes the industry to which the related company itself belongs.

[0051] S70: Determine the industry to which the enterprise to be classified belongs from several alternative industry categories based on information about related enterprises; Specifically, after identifying the related companies of the company to be classified and obtaining their related company information, the related companies are first classified based on their related company relationship information, identifying the related companies in the same industry corresponding to multiple candidate industries of the company to be classified. Then, a second industry score is calculated for a candidate industry based on the related company information of the same industry related companies in that candidate industry. At the same time, a first industry score is determined for the candidate industry based on its similarity to the company to be classified. The first and second industry scores are weighted and calculated to finally determine the comprehensive industry score for the candidate industry. For each candidate industry of the company to be classified, there is a corresponding comprehensive industry score. The candidate industry with the highest comprehensive industry score is selected as the industry of the company to be classified, thus completing the industry classification of the company to be classified.

[0052] Reference Figure 2 , Figure 2This application describes a specific possible embodiment. For a company to be classified, after the first similarity threshold screening, there are three candidate industries: Industry A, Industry B, and Industry C. For Industry A, the related companies of the company to be classified are searched, and there are three related companies in the same industry: Industry A, Industry B, and Industry C. Based on the related company information of these three related companies, a second industry score for Industry A is calculated. Simultaneously, the similarity between Industry A and the company to be classified is used as the first industry score for Industry A. The first and second industry scores are calculated using preset first and second weighting coefficients to finally determine the comprehensive industry score for Industry A. Following the same steps, the comprehensive industry scores for Industry B and Industry C are obtained respectively. The industry with the highest comprehensive industry score is selected as the industry to be classified for the company.

[0053] When calculating the second industry score, the information of related companies in the same industry is specifically considered. The more related companies there are in the same industry for a candidate company and the closer the relationship between each related company and the company to be classified, the greater the likelihood that the company to be classified belongs to that candidate industry. The corresponding performance in the second industry score is a higher score.

[0054] The degree of association between related companies in the same industry and the company to be classified can be inferred from the related company information of the related companies in the same industry. It should be noted that different related companies in the same industry may have different relationships with the company to be classified. In this case, the method of determining the degree of association between related companies in the same industry and the company to be classified will be different.

[0055] For example, regarding company A in the same industry, based on the related enterprise relationship information of company A, it can be known that company A has a cooperative relationship with the company to be classified. Therefore, the degree of correlation between company A and the company to be classified can be evaluated based on the evaluation criteria for the degree of cooperative relationship. It should be noted that the evaluation criteria for the degree of cooperative relationship are derived from expert analysis in the relevant field based on expert knowledge. In one feasible embodiment of this application, the evaluation criteria for the degree of cooperative relationship considers three indicators: the length of cooperation between company A and the company to be classified, the amount of cooperative transactions, and the number of cooperative projects. Based on these three indicators, the degree of correlation between company A and the company to be classified is determined, and a correlation score is obtained. Based on the correlation scores of all related enterprises in a candidate industry, a second industry score can be calculated.

[0056] Reference Figure 3This application also provides an enterprise industry classification system based on big data, which specifically includes the following modules: The enterprise information acquisition module 301 is used to acquire enterprise information of enterprises to be classified. The enterprise keyword extraction module 302 is used to extract enterprise keywords of the enterprise to be classified based on enterprise information, and to determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; The similarity calculation module 303 is used to calculate the similarity between the enterprise keyword set and the pre-set keyword sets of each industry using a pre-set similarity calculation model. The alternative industry determination module 304 is used to select the industry categories corresponding to the set of industry keywords with similarity greater than the similarity threshold as alternative industries for the enterprise to be classified.

[0057] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0058] This application also discloses an electronic device 400. (See reference...) Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device 400 disclosed in an embodiment of this application. The electronic device 400 may include: at least one processor 401, at least one network interface 404, a user interface 403, a memory 405, and at least one communication bus 402.

[0059] The communication bus 402 is used to enable communication between these components.

[0060] The user interface 403 may include a display screen and a camera. Optionally, the user interface 403 may also include a standard wired interface and a wireless interface.

[0061] The network interface 404 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0062] The processor 401 may include one or more processing cores. The processor 401 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 405, and by calling data stored in memory 405. Optionally, the processor 401 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 401 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip without being integrated into the processor 401.

[0063] The memory 405 may include random access memory (RAM) or read-only memory. Optionally, the memory 405 may include a non-transitory computer-readable storage medium. The memory 405 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 405 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 405 may also be at least one storage device located remotely from the aforementioned processor 401. (Refer to...) Figure 4 The memory 405, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program based on a big data-driven enterprise industry classification method.

[0064] exist Figure 4In the illustrated electronic device 400, the user interface 403 is mainly used to provide an input interface for the user and acquire user input data; while the processor 401 can be used to call an application stored in the memory 405, which is a big data-based enterprise industry classification method. When executed by one or more processors 401, the electronic device 400 performs one or more of the methods described in the above embodiments. It should be noted that, for the foregoing method embodiments, for the sake of simplicity, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0065] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0066] In the various embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between apparatuses or units may be electrical or other forms.

[0067] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0068] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0069] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device 405. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage device 405 and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage device 405 includes various media capable of storing program code, such as a USB flash drive, external hard drive, magnetic disk, or optical disk.

[0070] The above description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the specification and the disclosure of practical truths.

[0071] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope of this disclosure is defined by the claims.

Claims

1. A method for classifying enterprise industries based on associated enterprise information and the BERT model, characterized in that, The method includes the following steps: Obtain enterprise information for companies to be classified; Extract enterprise keywords from the enterprise information to be classified, and determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; The enterprise keyword set and the industry keyword set are encoded by a pre-trained BERT model to obtain the enterprise keyword set vector and the industry keyword set vector, respectively. Then, the cosine similarity between the enterprise keyword set vector and the industry keyword set vector is calculated. The industry categories corresponding to the set of industry keywords with similarity greater than the similarity threshold are selected as the candidate industries to which the enterprise to be classified belongs; If the candidate industries of the enterprise to be classified are not unique, then the related enterprises of the enterprise to be classified are determined, and the related enterprise information of each related enterprise is obtained. The related enterprise information includes related enterprise relationship information and related enterprise industry classification information. The similarity between the related enterprise information and each of the candidate industries is processed according to the preset comprehensive industry scoring calculation rules, and the comprehensive industry score of each of the candidate industries is calculated. The candidate industry with the highest comprehensive industry score is selected as the industry to which the enterprise to be classified belongs.

2. The enterprise industry classification method based on associated enterprise information and the BERT model according to claim 1, in processing the similarity between the associated enterprise information and each of the candidate industries according to the preset comprehensive industry scoring calculation rules, and calculating the comprehensive industry score of each of the candidate industries, specifically includes: For a candidate industry, the similarity between the candidate industry and the enterprise to be classified is used as the first industry score; Among the related enterprises of the enterprise to be classified, identify the related enterprises in the same industry category as the candidate enterprise; Calculate the second industry score based on the related enterprise information of the related enterprises in the same industry; The first industry score and the second industry score are weighted and calculated to complete the calculation of the comprehensive industry score of the candidate industry.

3. The enterprise industry classification method based on related enterprise information and the BERT model according to claim 1, characterized in that, Before obtaining the enterprise information of the enterprises to be classified, the method for creating an industry keyword set is also included. The method for creating an industry keyword set specifically includes: Obtain the National Economic Industry Classification document; Create multiple sets of industry keywords corresponding to the industry categories specified in the National Economic Industry Classification document; Obtain the industry keywords corresponding to the industry categories of each industry keyword set; The industry keywords for each industry category are stored in their respective industry keyword sets, thus completing the creation of each industry keyword set.

4. The enterprise industry classification method based on related enterprise information and the BERT model according to claim 3, characterized in that, The industry keywords for each industry keyword set corresponding to the industry category specifically include: For a given set of industry keywords, the first industry keyword is obtained based on the industry annotations of the industry categories corresponding to the set of industry keywords in the National Economic Industry Classification document. Obtain the enterprise information of companies in the same industry from a pre-set enterprise database, wherein the industry to which the companies in the same industry belong is the same as the industry category corresponding to the industry keyword set; Obtain second industry keywords based on the company information of the companies in the same industry.

5. The enterprise industry classification method based on related enterprise information and the BERT model according to claim 2, characterized in that, The calculation of the second industry score based on the related enterprise information of the related enterprises in the same industry specifically includes: Based on the association relationship information of each of the associated enterprises, each of the associated enterprises is classified, and multiple candidate industries of the enterprise to be classified are identified as related enterprises in the same industry. The second industry score of the candidate industry is calculated based on the related enterprise information of the related enterprises of the same industry in the candidate industry.

6. The enterprise industry classification method based on related enterprise information and the BERT model according to claim 5, characterized in that, The calculation of the second industry score for the candidate industry based on the related enterprise information of the related enterprises in the same industry as the candidate industry specifically includes: Calculate the correlation score between the related companies in the same industry and the company to be classified; The second industry score is calculated based on the correlation scores of all related companies in the same industry as the candidate industry.

7. The enterprise industry classification method based on associated enterprise information and the BERT model according to claim 6, characterized in that, The calculation of the correlation score between the related enterprises in the same industry and the enterprise to be classified specifically includes: Based on the preset evaluation criteria for the degree of cooperation, the degree of correlation between the related enterprises in the same industry and the enterprises to be classified is determined in terms of the years of cooperation, the amount of cooperation transactions, and the number of cooperation projects, so as to obtain the degree of correlation score.

8. An enterprise industry classification system based on associated enterprise information and the BERT model, characterized in that, The system includes: The enterprise information acquisition module is used to acquire enterprise information of enterprises to be classified. The enterprise keyword extraction module is used to extract enterprise keywords of the enterprise to be classified based on the enterprise information, and to determine the enterprise keyword set of the enterprise to be classified based on the enterprise keywords; The similarity calculation module is used to encode the enterprise keyword set and the industry keyword set through a pre-trained BERT model, to obtain the enterprise keyword set vector and the industry keyword set vector respectively, and then calculate the cosine similarity between the enterprise keyword set vector and the industry keyword set vector. The alternative industry determination module is used to select the industry category corresponding to the set of industry keywords with a similarity greater than a similarity threshold as the alternative industry of the enterprise to be classified. The associated enterprise determination module is used to determine the associated enterprises of the enterprise to be classified if the candidate industries of the enterprise to be classified are not unique, and to obtain the associated enterprise information of each associated enterprise. The associated enterprise information includes associated enterprise relationship information and associated enterprise industry classification information. The comprehensive scoring calculation module is used to process the similarity between the related enterprise information and each of the candidate industries according to the preset comprehensive industry scoring calculation rules, and calculate the comprehensive industry score of each of the candidate industries. The industry determination module is used to select the candidate industry with the highest comprehensive industry score as the industry to which the enterprise to be classified belongs.

9. An electronic device, characterized in that, The device includes a processor (401), a memory (405), a user interface (403), and a network interface (404). The memory (405) is used to store instructions. The user interface (403) and the network interface (404) are used to communicate with other devices. The processor (401) is used to execute the instructions stored in the memory (405) to cause the electronic device (400) to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, perform the steps of the method as described in any one of claims 1-7.

Images