Enterprise industry classification method and system based on various enterprise data analysis

Abstract

The invention belongs to the field of enterprise industry analysis, and provides an enterprise industry classification method and system based on multi-enterprise data analysis, and the method comprises the steps: carrying out the preprocessing of the obtained enterprise data, and obtaining the enterprise information; generating an enterprise static data industry vector from the enterprise information by using a pre-trained basic information model; generating an enterprise dynamic data industry vector from the enterprise information based on a pre-trained operation data item model; splicing the enterprise static data industry vector and the enterprise dynamic data industry vector into a comprehensive enterprise data industry vector, and reasoning the comprehensive enterprise data industry vector by using a pre-trained industry comprehensive model to obtain an enterprise industry vector; and performing similarity comparison on the enterprise industry vector and the industry embedding vector in the industry embedding vector library to confirm the industry to which the enterprise belongs. According to the invention, the problem that the industry of the existing enterprise data cannot be accurately marked is solved, and the marking accuracy is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of enterprise industry analysis, and specifically relates to an enterprise industry classification method and system based on multiple enterprise data analysis. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The goal of industry analysis is to gain a deeper understanding of the internal relevance and structural characteristics of an industry. It not only focuses on the development of a single industry, but also involves the overall situation of multiple related industries. The results of industry analysis include industrial planning, industrial policies, industrial chain maps, etc., which mainly serve industry managers and researchers; industry analysis not only includes the meaning of the industry, but also involves the organization and structure of a wider range of economic activities. Industry can refer to a series of interrelated industry clusters that may share raw materials, technological bases, market channels or upstream and downstream production relationships. For example, the automobile industry not only includes automobile manufacturing, but also involves multiple related industries such as parts supply, sales services, and repair and maintenance.

[0004] In the existing enterprise industry classification method, a knowledge base is constructed and prompt words are generated using prompt word templates, and the enterprise industry chain is generated by a large model based on the prompt words. The disadvantage of this method is that the data used to construct the knowledge base are mainly static definitions and attribute data, and do not contain dynamic data such as higher-value business data; there is no targeted training for the large model, and the accuracy is difficult to guarantee; there is also a technical solution that classifies the enterprise's business data, and then uses a preliminary model to group and analyze the industry chain, and then uses a joint model to refine the preliminary model results, and finally comprehensively analyzes the joint model results; however, the defect of this method is that in the process of multi-level model transmission, the results are directly screened, while other implicit content in the data is ignored; and the necessary data merging work is not done, resulting in related data being divided into different models, thereby reducing the accuracy of enterprise industry classification. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes an enterprise industry classification method and system based on the analysis of multiple enterprise data. The present invention comprehensively considers enterprise static data and enterprise dynamic data for data merging, uses different data to train corresponding models, and uses the trained models to classify enterprise industries, thereby improving the classification accuracy and solving the problem that the existing enterprise data cannot accurately mark the industry to which it belongs.

[0006] According to some embodiments, a first solution of the present invention provides a method for enterprise industry classification based on analysis of multiple enterprise data, using the following technical solutions:

[0007] An enterprise industry classification method based on the analysis of various enterprise data, comprising:

[0008] Preprocess the obtained enterprise data to obtain enterprise information;

[0009] Use the pre-trained basic information model to generate an enterprise static data industry vector from the enterprise information;

[0010] Based on the pre-trained business data item model, generate an enterprise dynamic data industry vector from the enterprise information;

[0011] Concatenate the enterprise static data industry vector and the enterprise dynamic data industry vector into a comprehensive enterprise data industry vector, and use the pre-trained industry comprehensive model to reason about the comprehensive enterprise data industry vector to obtain an enterprise industry vector;

[0012] Use the enterprise industry vector to compare the similarity with the industry embedding vectors in the industry embedding vector library to confirm the industry to which the enterprise belongs.

[0013] Further, the enterprise information includes enterprise static data and enterprise dynamic data;

[0014] Among them, the enterprise static data is a piece of data composed of enterprise name, enterprise industry classification, enterprise business scope, enterprise trademark, enterprise patent, enterprise website, enterprise online store, enterprise products and enterprise services;

[0015] The enterprise dynamic data is multiple business data items composed of bidding data, recruitment data and news public opinion.

[0016] Further, the training of the basic information model is carried out using the enterprise static data with pre-labeled industry classification;

[0017] The training of the business data item model is carried out using the enterprise dynamic data with pre-labeled industry classification;

[0018] When the training of the basic information model and the business data item model converges, according to the enterprise static data industry vector trained by the basic information model and the enterprise dynamic data industry vector trained by the business data item model, use the pre-labeled enterprise industry data to train the enterprise comprehensive model;

[0019] Among them, the enterprise static data corresponds to a basic information model, and the number of business data item models is determined according to the number of data items in the enterprise dynamic data.

[0020] Further, the enterprise dynamic data industry vector is the mean value of the industry vectors output by multiple business data item models.

[0021] Further, the basic information model, the business data item model, and the industrial comprehensive model are all regression models for generating embedding vectors.

[0022] Further, the industrial category of an enterprise is confirmed by comparing the similarity between the enterprise's industrial vector and the industrial embedding vectors in the industrial embedding vector library, specifically as follows:

[0023] Calculate the similarity between the enterprise's industrial vector and the industrial embedding vectors in the industrial embedding vector library;

[0024] If the similarity exceeds the set threshold, the industrial category of the enterprise is determined based on this industrial embedding vector.

[0025] According to some embodiments, the second solution of the present invention provides an enterprise industrial classification system based on the analysis of multiple enterprise data, adopting the following technical solution:

[0026] An enterprise industrial classification system based on the analysis of multiple enterprise data, comprising:

[0027] An enterprise data processing module, configured to preprocess the acquired enterprise data to obtain enterprise information;

[0028] An enterprise static data processing module, configured to generate an enterprise static data industrial vector from the enterprise information by using a pre-trained basic information model;

[0029] An enterprise dynamic data processing module, configured to generate an enterprise dynamic data industrial vector from the enterprise information based on a pre-trained business data item model;

[0030] An enterprise industrial vector determination module, configured to splice the enterprise static data industrial vector and the enterprise dynamic data industrial vector into a comprehensive enterprise data industrial vector, and perform inference on the comprehensive enterprise data industrial vector by using a pre-trained industrial comprehensive model to obtain an enterprise industrial vector;

[0031] An enterprise industrial category determination module, configured to compare the similarity between the enterprise industrial vector and the industrial embedding vectors in the industrial embedding vector library to confirm the industrial category of the enterprise.

[0032] A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the steps in an enterprise industrial classification method based on the analysis of multiple enterprise data as described in the first aspect above.

[0033] According to some embodiments, the fourth solution of the present invention provides a computer device.

[0034] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in a method for enterprise industry classification based on analysis of multiple enterprise data as described in the first aspect above are implemented.

[0035] According to some embodiments, a fifth aspect of the present invention provides a computer program product or a computer program.

[0036] The present invention provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in the enterprise industry classification method based on analysis of multiple enterprise data as described in the first aspect above.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention comprehensively considers enterprise static data and enterprise dynamic data to merge data, uses different data to train corresponding models, and uses the trained models to classify enterprise industries, thereby improving classification accuracy and solving the problem that existing enterprise data cannot accurately mark the industry to which they belong. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0040] Figure 1 This is a flow chart of a method for enterprise industry classification based on analysis of multiple enterprise data in an embodiment of the present invention;

[0041] Figure 2 is a flow chart of enterprise industry classification in an embodiment of the present invention;

[0042] Figure 3 This is a flow chart of model training in an embodiment of the present invention;. DETAILED DESCRIPTION

[0043] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0045] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they specify the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0046] In the case of no conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0047] Embodiment 1

[0048] This embodiment provides a method for classifying enterprise industries based on the analysis of various enterprise data. This embodiment takes the application of this method to a server as an example. It can be understood that this method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is realized through the interaction between the terminal and the server. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, web servers, cloud communications, middleware services, domain name services, security services CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server can be directly or indirectly connected through wired or wireless communication methods, and this application does not make any restrictions here. In this embodiment, the method includes the following steps:

[0049] Preprocess the obtained enterprise data to obtain enterprise information;

[0050] Use the pre-trained basic information model to generate an enterprise static data industry vector from the enterprise information;

[0051] Based on the pre-trained business data item model, generate an enterprise dynamic data industry vector from the enterprise information;

[0052] Concatenate the enterprise static data industry vector and the enterprise dynamic data industry vector into a comprehensive enterprise data industry vector, and use the pre-trained industry comprehensive model to reason about the comprehensive enterprise data industry vector to obtain an enterprise industry vector;

[0053] Use the enterprise industry vector to compare the similarity with the industry embedding vectors in the industry embedding vector library to confirm the industry to which the enterprise belongs.

[0054] As Figure 2 shown, the specific process of the method described in this embodiment is:

[0055] 1. Retrieve enterprise dynamic data item by item from the data source - enterprise operation data for inference calculation. This includes: data item model selection, data vectorization, etc.

[0056] 2. Use the enterprise operation data item model to obtain the inference result of enterprise dynamic data - the industrial vector of enterprise dynamic data, and save the inference result of enterprise dynamic data to the enterprise operation data item database. This storage is used to calculate the average vector of the inference results of enterprise operation data items.

[0057] 3. Calculate the average vector of the inference results of enterprise dynamic data for use in the inference calculation of the industrial comprehensive model.

[0058] 4. Retrieve enterprise static data from the data source for inference calculation.

[0059] 5. Use the basic information model to obtain the inference result of enterprise static data - the industrial vector of enterprise static data for use in the inference calculation of the industrial comprehensive model.

[0060] 6. Based on the vector of the inference result of enterprise static data and the average vector of the inference results of enterprise dynamic data, splice them into a comprehensive enterprise data industrial vector, and use the industrial comprehensive model to infer the enterprise industrial vector.

[0061] 7. Use the industrial comprehensive model to infer the enterprise industrial vector, query the industrial vector with high similarity in the vector database - the industrial embedding vector library, which is the enterprise industry. Use the enterprise industrial vector inferred by the industrial comprehensive model to query the similar industrial vector in the industrial embedding vector library, and the industrial vector with a similarity greater than the threshold is the enterprise's industrial vector.

[0062] It should be noted that the technical solution proposed in this embodiment is not limited to specific algorithms, models, nor to computing engines, databases, etc. As long as the algorithms, models, computing engines, databases, etc. that meet the conditions can be used to implement this solution.

[0063] In a specific embodiment, as Figure 3 shown, the detailed process of model design and training in this method specifically includes:

[0064] 1. Classification of enterprise data. Enterprise data is generally divided into two categories: static data representing the enterprise status and dynamic data representing the enterprise behavior.

[0065] Static data represents the data of the enterprise status, including: enterprise name, enterprise industry classification, enterprise business scope, enterprise trademark, enterprise patent, enterprise website, enterprise online store, enterprise products, enterprise services, etc.;

[0066] Dynamic data represents the data of the enterprise behavior, including: bidding data, recruitment data, news and public opinion, etc.

[0067] Among them, the dynamic data (business information data) can be further divided according to specific data items. For example: bidding data, recruitment data, news and public opinion, etc. For the static data, they need to be combined together to avoid the instability of the model caused by the outliers of individual attributes. Whether it is dynamic data or static data, fields unrelated to the enterprise industry should be removed to reduce interference. The basic information data corresponds to a basic information model; for the business information data, there are models corresponding to the specific data items.

[0068] Among them, the dynamic data and the business information data refer to the same kind of data. It is called business information because these data are generated by the enterprise's business activities. It is called dynamic data because these data are composed of one behavior or a series of behaviors with a beginning and an end (for example: there are several enterprise recruitment information, each of which is independent; there are several bids for enterprise bidding information, and each bid is independent;), not every enterprise has only one copy.

[0069] The division of the enterprise's static and dynamic data is mainly based on whether the object described by the data is transactional. If the described object is transactional, then it is classified as dynamic data, otherwise it is static data. Extract the industry-related fields of the static data and use them to form a unified static data.

[0070] Thus, it can be seen that the static data, that is, the basic information data corresponds to a basic information model; for the business information data, there are models corresponding to the specific data items, and for the dynamic data, different models should be used according to the specific type, that is, each type of dynamic data has its corresponding model.

[0071] 2. Train the enterprise attribute embedding vector model. Enterprise attributes are often described by strings and need to be converted into embedding vectors to be used by the model. The existing embedding vectors are all in units of words, but enterprise attributes are often composed of multiple words (such as the enterprise's business scope, etc.), so an embedding vector model of its own needs to be trained for these enterprise attributes; the vectorization of enterprise data. Train the corresponding embedding vectors for each attribute of the enterprise data. Replace the attribute values with the corresponding embedding vectors to obtain the vectorized enterprise data. Enterprise attributes with inconsistent structures cannot be used as standard model input data. The enterprise attribute embedding vector model generates standard vector data that can be used for model input while retaining the characteristics of enterprise attributes. In actual analysis, it is used to convert the attribute data of the enterprise into vector data as the model input data.

[0072] 3. Data model design. The data models here are mainly divided into three categories: basic information model, business data item model, and industry comprehensive model.

[0073] Among them, the basic information model is a model that takes the static data of an enterprise as input, and each enterprise has only one piece of data; the business data item model is a model that takes the dynamic data item data corresponding to an enterprise as input, and each enterprise can have zero or multiple pieces of data; the industrial comprehensive model is a model that takes the enterprise inference result vector of the basic information model and the average value vector of the inference results of each business data item model as data together, and each enterprise has only one piece of data.

[0074] The basic information model, the business data item model, and the industrial comprehensive model are all regression models that generate embedding vectors.

[0075] The role of the basic information model is to infer the industrial vector (result vector) of an enterprise based on the basic information of the enterprise (enterprise static data); the role of the business data item model is to infer the industrial vector (result vector) of an enterprise based on the behavioral data of the enterprise (enterprise dynamic data); the role of the industrial comprehensive model is to synthesize the industrial vector of the enterprise inferred by the basic information model and the industrial vector of the enterprise inferred by the business data item model, and infer the industrial vector of the enterprise that combines the enterprise static data and the enterprise static data. This industrial vector is also the vector used to calculate the industry to which the enterprise belongs. Using the basic information model or the business data item model alone can also achieve enterprise industry classification, but the data range used is limited and the accuracy is low. Therefore, the industrial comprehensive model is used to combine the inference results of the basic information model and the business data item model to generate a more accurate enterprise industry classification.

[0076] According to the enterprise information, use the basic information model, the business data item model, and the industrial comprehensive model to obtain the enterprise industrial vector. The process is as follows: use the basic information model to generate the industrial vector of the enterprise static data; use the business data item model to generate multiple industrial vectors, and calculate the average value of these industrial vectors as the industrial vector of the enterprise dynamic data; concatenate the industrial vector of the enterprise static data and the industrial vector of the enterprise dynamic data into one piece of input data, and use the industrial comprehensive model to infer and obtain the enterprise industrial vector.

[0077] Use the enterprise industrial vector to query the industrial vector with high similarity in the industrial embedding vector library, which is the industry of the enterprise.

[0078] 4. Train the industrial embedding vector model. The total number of industries is not large, and it is easy to train the industrial embedding vector model. The self-trained industrial embedding vector model is relatively small and more dispersed, which can reduce misclassification.

[0079] Specifically, train the industrial embedding vector according to industrial relevance.

[0080] 5. Train the data model. Use the labeled data to first train the basic information model and the business data item model. After the basic information model and the business data item model converge, then train the industrial comprehensive model.

[0081] The static data model (basic information model) and each dynamic data model (business data item model) are trained independently, and the labeled data for training does not use whether an enterprise belongs to a certain industry as a label. Instead, according to the content of the training data, labels are assigned starting from specific data; the inference result of all models is an industry embedding vector.

[0082] Vectorization of enterprise data. For each attribute of enterprise data, a corresponding embedding vector is trained. By replacing the attribute value with the corresponding embedding vector, the vectorized enterprise data is obtained.

[0083] Labeled data during training. Data labeling is divided into two categories: a. The industry to which the enterprise belongs; b. The industry to which a single piece of enterprise data belongs.

[0084] Use the labeled data of a single piece of enterprise to train the basic information model and the business data item model. Use the labeled data of the industry to which the enterprise belongs to train the industry comprehensive model.

[0085] The comprehensive inference model (industry comprehensive model) is formed by combining the inference result vector of the static data model and the average vector of the inference result vectors of each dynamic data model (forming a piece of data with a length of twice the industry embedding vector, where the inference result vector of the static data model is in the front and the average vector of the inference result vectors of the dynamic data models is in the back) as the input data of the model, and use the comprehensive inference model to calculate the industry vector of the enterprise.

[0086] The training of the basic information model is carried out using the static enterprise data with pre-labeled industry classifications;

[0087] The training of the business data item model is carried out using the dynamic enterprise data with pre-labeled industry classifications;

[0088] When the training of the basic information model and the business data item model converges, according to the industry vectors of the enterprise static data trained by the basic information model and the industry vectors of the enterprise dynamic data trained by the business data item model, use the pre-labeled data of the industry to which the enterprise belongs to train the enterprise comprehensive model;

[0089] Among them, the enterprise static data corresponds to a basic information model, and the number of business data item models is determined according to the number of data items in the enterprise dynamic data; and the enterprise dynamic data industry vector is the average of the industry vectors output by multiple business data item models.

[0090] 6. Build an industry embedding vector library. Train the industry embedding vector model, build the industry embedding vector library, and import the trained industry embedding vectors into the vector database - the industry embedding vector library.

[0091] 7. Model evaluation. Use the test data to evaluate the trained model and determine whether the model meets the requirements.

[0092] Embodiment 2

[0093] This embodiment provides an enterprise industry classification system based on the analysis of various enterprise data, including:

[0094] An enterprise data processing module configured to preprocess the acquired enterprise data to obtain enterprise information;

[0095] An enterprise static data processing module configured to generate an enterprise static data industry vector from the enterprise information using a pre-trained basic information model;

[0096] An enterprise dynamic data processing module configured to generate an enterprise dynamic data industry vector from the enterprise information based on a pre-trained business data item model;

[0097] An enterprise industry vector determination module configured to concatenate the enterprise static data industry vector and the enterprise dynamic data industry vector into a comprehensive enterprise data industry vector, and use a pre-trained industry comprehensive model to infer the comprehensive enterprise data industry vector to obtain an enterprise industry vector;

[0098] An enterprise affiliated industry determination module configured to compare the similarity between the enterprise industry vector and the industry embedding vectors in the industry embedding vector library to confirm the enterprise affiliated industry.

[0099] The examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the content disclosed in the above Embodiment 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer executable instructions.

[0100] In the above embodiments, the descriptions of each embodiment have their own emphases. For parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0101] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the above module division is only a logical function division. In actual implementation, there can be other division methods. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0102] Embodiment 3

[0103] This embodiment provides a computer-readable storage medium with a computer program stored thereon. When the program is executed by a processor, it implements the steps in an enterprise industry classification method based on the analysis of various enterprise data as described in Embodiment 1 above.

[0104] Example 4

[0105] This embodiment provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps in an enterprise industry classification method based on multiple enterprise data analyses as described in Embodiment 1 above.

[0106] Example 5

[0107] This embodiment provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and these computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in an enterprise industry classification method based on multiple enterprise data analyses as described in Embodiment 1 above.

[0108] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of an embodiment implemented in hardware, a software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories and optical memories, etc.) containing computer-usable program code.

[0109] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0110] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0111] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the functions specified in one process or multiple processes and / or blocks Figure 1 one process or multiple processes and / or blocks Figure 1 or steps of the functions specified in multiple blocks.

[0112] Those of ordinary skill in the art can understand that all or part of the processes of the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disc, a read-only memory (ROM), or a random access memory (RAM), etc.

[0113] Although the specific implementation manners of the present invention have been described above in conjunction with the accompanying drawings, they are not limitations on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. A method for enterprise industry classification based on analysis of multiple enterprise data, characterized in that: include: Pre-process the acquired enterprise data to obtain enterprise information; Use the pre-trained basic information model to generate enterprise static data industry vectors from enterprise information; Generate enterprise dynamic data industry vectors from enterprise information based on pre-trained business data item models; The static data industry vector of the enterprise and the dynamic data industry vector of the enterprise are spliced ​​into a comprehensive enterprise data industry vector, and the pre-trained industry comprehensive model is used to infer the comprehensive enterprise data industry vector to obtain the enterprise industry vector; The similarity between the enterprise industry vector and the industry embedding vector in the industry embedding vector library is compared to confirm the industry to which the enterprise belongs.

2. The enterprise industry classification method based on multiple enterprise data analysis as claimed in claim 1, characterized in that: The enterprise information includes enterprise static data and enterprise dynamic data; Among them, enterprise static data is a piece of data consisting of enterprise name, enterprise industry classification, enterprise business scope, enterprise trademark, enterprise patent, enterprise website, enterprise online store, enterprise products and enterprise services; The enterprise dynamic data is a plurality of business data items consisting of bidding data, recruitment data and news and public opinion.

3. The enterprise industry classification method based on multiple enterprise data analysis as claimed in claim 1, characterized in that: The basic information model is trained using static enterprise data with pre-labeled industry classifications; The training of the business data item model is carried out using the enterprise dynamic data with pre-labeled industry classification; When the training of the basic information model and the business data item model converges, the enterprise comprehensive model is trained using the pre-labeled enterprise industry data according to the enterprise static data industry vector trained by the basic information model and the enterprise dynamic data industry vector trained by the business data item model; Among them, the enterprise static data corresponds to a basic information model, and the number of business data item models is determined according to the number of data items in the enterprise dynamic data.

4. The enterprise industry classification method based on multiple enterprise data analysis as claimed in claim 3, characterized in that: The enterprise dynamic data industry vector is the average of the industry vectors output by multiple business data item models.

5. The enterprise industry classification method based on multiple enterprise data analysis as claimed in claim 1, characterized in that: The basic information model, the business data item model and the industry comprehensive model are all regression models for generating embedded vectors.

6. The enterprise industry classification method based on multiple enterprise data analysis as claimed in claim 1, characterized in that: Use the enterprise industry vector to compare the similarity with the industry embedding vector in the industry embedding vector library to confirm the industry to which the enterprise belongs, specifically: Calculate the similarity between the enterprise industry vector and the industry embedding vector in the industry embedding vector library; If the similarity exceeds the set threshold, the industry to which the enterprise belongs is determined based on the industry embedding vector.

7. An enterprise industry classification system based on analysis of multiple enterprise data, characterized in that: include: The enterprise data processing module is configured to pre-process the acquired enterprise data to obtain enterprise information; The enterprise static data processing module is configured to generate an enterprise static data industry vector from enterprise information using a pre-trained basic information model; The enterprise dynamic data processing module is configured to generate an enterprise dynamic data industry vector from enterprise information based on a pre-trained business data item model; The enterprise industry vector determination module is configured to splice the enterprise static data industry vector and the enterprise dynamic data industry vector into a comprehensive enterprise data industry vector, and use a pre-trained industry comprehensive model to infer the comprehensive enterprise data industry vector to obtain the enterprise industry vector; The module for determining the industry to which an enterprise belongs is configured to compare the similarity between the enterprise industry vector and the industry embedding vector in the industry embedding vector library to confirm the industry to which the enterprise belongs.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a method for enterprise industry classification based on analysis of multiple enterprise data as described in any one of claims 1 to 6 are implemented.

9. A computer 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 steps in the enterprise industry classification method based on multiple enterprise data analysis as described in any one of claims 1-6 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the steps in the enterprise industry classification method based on multiple enterprise data analysis as described in any one of claims 1 to 6.

Images