Enterprise identification method and device, computer equipment and storage medium

Abstract

The embodiment of the invention belongs to the field of financial science and technology, and relates to an enterprise identification method and device, computer equipment and a storage medium, and the method comprises the following steps: obtaining enterprise multi-dimensional information; inputting the enterprise multi-dimensional information into a pre-trained feature extraction model for feature extraction to obtain enterprise multi-dimensional features; scoring the enterprise multi-dimensional features according to a preset scoring standard to obtain enterprise multi-dimensional scores; performing weighted calculation on the enterprise multi-dimensional score to obtain an enterprise multi-dimensional total score; judging whether the enterprise multi-dimensional total score is smaller than or equal to a preset threshold value or not; and if the enterprise multi-dimensional total score is smaller than or equal to the preset threshold value, identifying the enterprise corresponding to the enterprise multi-dimensional total score as an empty shell enterprise. According to the invention, accurate identification of empty-shell enterprises can be realized, so that normal development of financial businesses is effectively ensured.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, specifically to the field of financial technology, and in particular to an enterprise identification method, device, computer equipment and storage medium. Background Art

[0002] With the in-depth development of the market economy, enterprises, as the main body of economic activities, play an important role in promoting economic growth and creating employment opportunities. However, the existence of shell companies has brought a lot of harm to the economic environment. Shell companies, that is, companies that exist in form but lack normal business activities in essence, may be involved in illegal activities such as fraud, money laundering, and tax evasion. While seriously disrupting the market order, they also damage the environment of fair competition and affect the normal development of business in the financial industry.

[0003] In order to effectively identify shell companies, on-site inquiries are generally conducted through the relevant departments of the company. After actually inspecting the company, the actual information of the company is compared to whether it is consistent with the information registered with the industrial and commercial administration, or a judgment is made through one of the following situations: If the company lacks fixed assets, such as equipment, inventory, etc., it may be a sign of a shell company. At the same time, observe the company's personnel situation, including the number of employees, whether labor contracts are signed, whether social security is purchased, etc. Shell companies often lack actual employees or the number of employees does not match the scale of the business. Understand its actual business activities. If the company has not conducted business for a long time or its business activities are seriously inconsistent with its business scope, it may be a shell company.

[0004] The above identification method has the problems of being time-consuming, consuming too much manpower and having a long investigation cycle. Therefore, with the support of data from historical identification of real shell companies, the existing technology extracts the characteristics of historical real shell companies, and uses the characteristics to train and continuously iterate and optimize the model to obtain a method of accurately identifying a model for shell companies to identify whether a real company is a shell company. However, the above method of identifying shell companies is not accurate, and it is easy to make misjudgments, making it difficult to effectively identify shell companies, which is not conducive to the normal development of financial business. Summary of the invention

[0005] The purpose of the embodiments of the present application is to provide a method, apparatus, computer device and storage medium for enterprise identification to solve the problem of being unable to accurately and effectively identify shell companies.

[0006] In order to solve the above technical problems, the present application embodiment provides a method for enterprise identification, which adopts the following technical solution:

[0007] Obtain multi-dimensional information of enterprises;

[0008] Inputting the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain enterprise multi-dimensional features;

[0009] Scoring the multidimensional characteristics of the enterprise according to a preset scoring standard to obtain a multidimensional score of the enterprise;

[0010] Performing weighted calculation on the enterprise's multidimensional scores to obtain the enterprise's multidimensional total score;

[0011] Determine whether the enterprise's multi-dimensional total score is less than or equal to a preset threshold;

[0012] If the multi-dimensional total score of the enterprise is less than or equal to the preset threshold, the enterprise corresponding to the multi-dimensional total score of the enterprise is identified as a shell enterprise.

[0013] Furthermore, the step of obtaining multi-dimensional information of an enterprise specifically includes:

[0014] Acquire original enterprise data, perform data cleaning on the original enterprise data, and obtain enterprise standard data;

[0015] Input the enterprise standard data into the pre-trained classification model to obtain enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information;

[0016] The enterprise business location information, the enterprise asset form information, the enterprise personnel information, the enterprise business activity information, the enterprise business qualification information, and the enterprise risk information are associated with the corresponding enterprise to obtain the multi-dimensional information of the enterprise.

[0017] Furthermore, the step of inputting the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain the enterprise multi-dimensional features specifically includes:

[0018] Classifying the enterprise multidimensional information into variables to obtain enterprise multidimensional numerical variables and enterprise multidimensional categorical variables;

[0019] Normalizing the enterprise's multidimensional numerical variables to obtain the enterprise's multidimensional standard numerical values;

[0020] Performing dummy variable processing on the enterprise multidimensional category variables to obtain enterprise multidimensional category information;

[0021] Associating the enterprise multi-dimensional standard value with the enterprise multi-dimensional category information to obtain enterprise multi-dimensional valid information;

[0022] The enterprise's multi-dimensional valid information is input into the pre-trained feature extraction model to obtain the enterprise's multi-dimensional features.

[0023] Furthermore, the step of classifying the enterprise multidimensional information to obtain enterprise multidimensional numerical variables and enterprise multidimensional categorical variables specifically includes:

[0024] Acquiring multi-dimensional information of the enterprise;

[0025] Performing data cleaning and preprocessing on the enterprise multidimensional information to obtain standard enterprise multidimensional information;

[0026] Identify the variable type corresponding to the standard enterprise multi-dimensional information, wherein the variable type includes a numerical variable and a categorical variable;

[0027] The standard enterprise multidimensional information is classified according to the numerical variables and the categorical variables to obtain the enterprise multidimensional numerical variables and the enterprise multidimensional categorical variables.

[0028] Furthermore, before the step of inputting the enterprise multi-dimensional information into the pre-trained feature extraction model to extract features and obtain the enterprise multi-dimensional features, the following steps are also included:

[0029] Acquire multidimensional information of sample enterprises, and input the multidimensional information of sample enterprises into a preset model to obtain multidimensional features of the sample enterprises;

[0030] Calculate the multidimensional total score of the sample enterprises according to the multidimensional characteristics of the sample enterprises;

[0031] Determine whether the multi-dimensional total score of the sample enterprise is less than or equal to a preset threshold;

[0032] If the multi-dimensional total score of the sample enterprise is less than or equal to the preset threshold, the sample enterprise corresponding to the multi-dimensional total score of the sample enterprise is identified as a sample shell enterprise;

[0033] Obtaining historical shell company information, and judging whether the sample shell company is a real shell company based on the historical shell company information;

[0034] If the sample shell enterprise does not belong to the real shell enterprise, the preset model is adjusted, and the multidimensional features of the new sample enterprise are re-acquired to calculate the multidimensional total score of the new sample enterprise, and the new sample shell enterprise is re-judged according to the multidimensional total score of the sample enterprise, and then the new sample shell enterprise is re-judged according to the historical shell enterprise information whether it belongs to the real shell enterprise, until the new sample shell enterprise belongs to the real shell enterprise;

[0035] After the new sample shell company belongs to the real shell company, the preset model is used as the pre-trained feature extraction model.

[0036] Furthermore, the step of scoring the enterprise multidimensional characteristics according to the preset scoring criteria to obtain the enterprise multidimensional score specifically includes:

[0037] Obtain the enterprise category identifier corresponding to the enterprise's multi-dimensional features;

[0038] Extracting the preset scoring criteria from a database according to the enterprise category identifier;

[0039] The multi-dimensional characteristics of the enterprise are mapped to the preset scoring criteria to obtain the multi-dimensional score of the enterprise.

[0040] Furthermore, the enterprise multi-dimensional score includes the enterprise business location score, enterprise asset form score, enterprise personnel score, enterprise business activity score, enterprise business qualification score, and enterprise risk score. The step of weighted calculation of the enterprise multi-dimensional score to obtain the enterprise multi-dimensional total score specifically includes:

[0041] Obtain the score weight corresponding to the multi-dimensional score of the enterprise;

[0042] Weighting the enterprise business location score, the enterprise asset form score, the enterprise personnel score, the enterprise business activity score, the enterprise business qualification score, and the enterprise risk score according to the score weight;

[0043] The weighted scores of the enterprise's business location, asset form, personnel, business activities, business qualifications and risk are summed to obtain the enterprise's multi-dimensional total score.

[0044] In order to solve the above technical problems, the embodiment of the present application also provides a corporate identification device, which adopts the following technical solution:

[0045] Information acquisition module, used to obtain multi-dimensional information of enterprises;

[0046] A feature acquisition module is used to input the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain enterprise multi-dimensional features;

[0047] A feature scoring module is used to score the enterprise's multidimensional features according to a preset scoring standard to obtain a multidimensional score of the enterprise;

[0048] A scoring calculation module, used to perform weighted calculation on the enterprise's multi-dimensional scores to obtain the enterprise's multi-dimensional total score;

[0049] A total score judgment module is used to judge whether the multi-dimensional total score of the enterprise is less than or equal to a preset threshold;

[0050] The result identification module is used to identify the enterprise corresponding to the multi-dimensional total score of the enterprise as a shell enterprise if the multi-dimensional total score of the enterprise is less than or equal to the preset threshold.

[0051] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0052] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of any one of the above enterprise identification methods when executing the computer-readable instructions.

[0053] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0054] A computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the enterprise identification method described in any one of the above items are implemented.

[0055] Compared with the prior art, the embodiment of the present application has the following beneficial effects: the embodiment obtains enterprise multidimensional information; inputs the enterprise multidimensional information into the pre-trained feature extraction model for feature extraction to obtain enterprise multidimensional features; scores the enterprise multidimensional features according to the preset scoring criteria to obtain the enterprise multidimensional score; performs weighted calculation on the enterprise multidimensional score to obtain the enterprise multidimensional total score; determines whether the enterprise multidimensional total score is less than or equal to the preset threshold; if the enterprise multidimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a shell enterprise; if the enterprise multidimensional total score is greater than the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a non-shell enterprise. Through the above method, the embodiment can extract features based on multidimensional data related to the enterprise, and perform weighted scoring on each feature, so as to obtain the enterprise multidimensional total score reflecting the enterprise risk status, and effectively identify shell enterprises based on the enterprise multidimensional total score, thereby completing the accurate identification of shell enterprises, reducing the risk in the process of financial business development, and effectively ensuring the normal development of financial business. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0057] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0058] Figure 2 A flowchart of an embodiment of the enterprise identification method according to the present application;

[0059] Figure 3 yes Figure 2 A flowchart of a specific implementation of step S10;

[0060] Figure 4 yes Figure 2 A flowchart of a specific implementation of step S20;

[0061] Figure 5 yes Figure 4 A flowchart of a specific implementation of step S201;

[0062] Figure 6 yes Figure 2 A flowchart of a specific implementation of step S30;

[0063] Figure 7 yes Figure 2 A flowchart of a specific implementation of step S40;

[0064] Figure 8 is a schematic structural diagram of an embodiment of a corporate identification device according to the present application;

[0065] Fig. 9 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0067] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor does it refer to non-related or alternative embodiments that are mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0068] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0069] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0070] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0071] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0072] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0073] It should be noted that the enterprise identification method provided in the embodiment of the present application is generally executed by a server, and accordingly, the enterprise identification device is generally set in the server.

[0074] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0075] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for enterprise identification according to the present application. The enterprise identification method comprises the following steps:

[0076] Step S10, obtaining multi-dimensional information of the enterprise;

[0077] In this embodiment, enterprise multi-dimensional information refers to multi-dimensional information data for enterprises, and enterprise multi-dimensional information includes enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information. Among them, enterprise business location information includes the enterprise's registered address, main office location, geographical location of branches, etc., which is helpful to understand the market coverage and geographical layout of the enterprise; enterprise asset form information involves the asset structure of the enterprise, including fixed assets, current assets, intangible assets, etc., as well as the distribution and use of assets; enterprise personnel information includes the number, structure, educational background, professional skills, etc. of enterprise employees, which can reflect the human resource status and team capabilities of the enterprise; enterprise business activity information involves the main business, products or services, market positioning, customer relations, supply chain management, etc. of the enterprise, which is helpful to evaluate the market competitiveness and operating efficiency of the enterprise; enterprise business qualification information includes the enterprise's business license, industry qualifications, certification certificates, etc., which prove the legal right and professional ability of the enterprise in a specific field; enterprise risk information includes the enterprise's legal litigation record, debt situation, business risks, etc., which is helpful to evaluate the stability and potential risks of the enterprise.

[0078] Step S20, inputting the enterprise multi-dimensional information into a pre-trained feature extraction model to perform feature extraction to obtain enterprise multi-dimensional features;

[0079] In this embodiment, the pre-trained feature extraction model adopts the LightGBM model. The LightGBM model uses a decision tree-based learning algorithm to divide the continuous values ​​of the enterprise's multidimensional information into discrete boxes, and then use the boxes to construct a feature histogram during the training process, and then extract the corresponding enterprise multidimensional features based on the feature histogram. The continuous value of the above-mentioned enterprise multidimensional information refers to data information with a continuous nature, such as enterprise financial information, including income and expenditure, profit and loss, where income and expenditure are the company's monthly, quarterly or annual sales revenue, cost expenditure and other financial data, and profit and loss are the company's net profit, gross profit margin, net profit margin and other financial indicators for each time period. By dividing the continuous values ​​of the enterprise's multidimensional information into discrete directions, the model data is effectively simplified, the algorithm efficiency is improved, and the generalization ability of the model is improved.

[0080] Step S30, scoring the enterprise multidimensional characteristics according to a preset scoring standard to obtain an enterprise multidimensional score;

[0081] In this embodiment, the preset scoring standard can be an evaluation standard formulated based on historical enterprise data, which is obtained by inputting the historical enterprise data into a preset evaluation model and performing a weight analysis. The evaluation model can use a random forest or gradient boosting tree to use a feature selection algorithm to automatically select important features in the historical enterprise data, and assign corresponding weights according to the importance of the important features, thereby obtaining an evaluation standard, and using the evaluation standard as the preset scoring standard. The above scoring standard can be a quantitative scoring value or a qualitative grade.

[0082] Step S40, performing weighted calculation on the enterprise multi-dimensional score to obtain the enterprise multi-dimensional total score;

[0083] In this embodiment, the enterprise multidimensional score includes the enterprise business site score, enterprise asset form score, enterprise personnel score, enterprise business activity score, enterprise business qualification score, and enterprise risk score corresponding to the above-mentioned enterprise multidimensional information. The weighted calculation refers to the corresponding score calculation according to the different score weights set for the enterprise business site score, enterprise asset form score, enterprise personnel score, enterprise business activity score, enterprise business qualification score, and enterprise risk score. For example, the score weight corresponding to the enterprise business site score is 0.3, then the enterprise business site score = enterprise business site score × 0.3, the weight corresponding to the enterprise asset form score is 0.2, then the enterprise asset form score = enterprise asset form score × 0.2, after the above weighted calculation, the weighted calculation score corresponding to each score is obtained, and then the weighted calculation score corresponding to each score is summed up to obtain the enterprise multidimensional total score. The above-mentioned scoring weights can be obtained by inputting the historical enterprise data into the model for analysis.

[0084] Step S50, determining whether the enterprise multi-dimensional total score is less than or equal to a preset threshold;

[0085] In this embodiment, the upper limit of the enterprise multidimensional total score is set to 100 points. The lower the score, the higher the probability that the enterprise corresponding to the enterprise multidimensional total score is a shell enterprise. The preset threshold is set to 40 points. When the enterprise multidimensional total score is less than or equal to 40 points, it is judged that the enterprise corresponding to the enterprise multidimensional total score is a shell enterprise. The upper limit of the enterprise multidimensional total score and the preset threshold can be adjusted accordingly according to the actual situation. At the same time, the percentage of the preset threshold to the upper limit of the enterprise multidimensional total score can also be adjusted according to the actual situation. For example, in this embodiment, the percentage of the preset threshold to the upper limit of the enterprise multidimensional total score is 40%, which can be adjusted to 35%, then the preset threshold is 35 points, or adjusted to 30%, then the preset threshold is 30 points. The corresponding adjustment is made according to the actual situation to obtain a more accurate and effective judgment result.

[0086] In this embodiment, the adjustment of the preset threshold can be achieved through the following steps: data collection: collect multi-dimensional total score data of a large number of enterprises and labels of whether they are shell enterprises (obtained through manual review, external data sources, etc.); data preprocessing: clean and standardize the data to ensure data quality; set an initial threshold: set an initial preset threshold (such as 40 points) based on business experience or preliminary analysis; establish a classifier: use machine learning algorithms such as logistic regression, decision tree, random forest, etc., and train a binary classification model with the multi-dimensional total score of the enterprise as a feature to predict whether the enterprise is a shell enterprise; model evaluation: use the validation set to evaluate the model performance, including accuracy, recall, F 1 score and other indicators; Adjustment based on performance indicators: If the model's accuracy or F1 score in identifying shell companies is low, consider adjusting the preset threshold. For example, if the model tends to misjudge too many companies as shell companies (too many false positives), the threshold can be appropriately increased; conversely, if there are too many missed detections (too many false negatives), the threshold can be lowered; Dynamic threshold: Consider introducing a dynamic threshold adjustment mechanism, such as automatically adjusting the threshold based on factors such as the number of company registrations, industry characteristics, and policy changes in a specific time period; Iterative training: Retrain or fine-tune the model using the adjusted threshold, and continuously monitor model performance; Real-time monitoring: Monitor model performance in real time in a production environment, including key indicators such as accuracy, false positive rate, and false negative rate; Regular optimization: Regularly re-evaluate and adjust preset thresholds and model parameters based on monitoring results and new business data; System integration: Integrate the model into the company's existing financial risk control system to achieve real-time identification and monitoring of shell companies. In the specific implementation, manual methods can also be used to make simple adaptive adjustments. Manual review: Regular manual review of model prediction results, especially those near the threshold; expert opinion integration: combining expert opinions with model prediction results to further optimize thresholds and model parameters.

[0087] Step S60, if the enterprise multi-dimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multi-dimensional total score is identified as a shell enterprise;

[0088] In this embodiment, shell companies refer to companies that exist in name but have no business activities. Shell companies are usually registered to meet special purposes such as evading supervision, tax avoidance, money laundering, etc., or to conduct fraudulent activities, such as deceiving investors or obtaining loans. The existence of shell companies will bring certain harm and risks to the financial industry. This embodiment can identify shell companies and obtain an untrustworthy blacklist to effectively reduce risks in financial business.

[0089] Step S70: If the multi-dimensional total score of the enterprise is greater than the preset threshold, the enterprise corresponding to the multi-dimensional total score of the enterprise is identified as a non-shell enterprise.

[0090] In this embodiment, a non-shell enterprise refers to an enterprise unit that is registered and established through normal channels and has fixed operating personnel, normal business scope and normal business transactions. This embodiment can obtain a trustworthy white list by identifying non-shell enterprises to improve the security of financial business during the financial business.

[0091] This embodiment obtains enterprise multidimensional information; inputs the enterprise multidimensional information into a pre-trained feature extraction model for feature extraction to obtain enterprise multidimensional features; scores the enterprise multidimensional features according to a preset scoring standard to obtain an enterprise multidimensional score; performs weighted calculation on the enterprise multidimensional score to obtain an enterprise multidimensional total score; determines whether the enterprise multidimensional total score is less than or equal to a preset threshold; if the enterprise multidimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a shell enterprise; if the enterprise multidimensional total score is greater than the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a non-shell enterprise. Through the above method, this embodiment can extract features based on multidimensional data related to the enterprise, and perform weighted scoring on each feature, thereby obtaining an enterprise multidimensional total score that reflects the risk status of the enterprise, and effectively identifies shell enterprises based on the enterprise multidimensional total score, thereby completing the accurate identification of shell enterprises, reducing the risks in the process of carrying out financial business, and effectively ensuring the normal development of financial business.

[0092] Continue to refer Figure 3 In some optional implementations of this embodiment, step S10 includes the following steps:

[0093] Step S101, obtaining original enterprise data, and performing data cleaning on the original enterprise data to obtain enterprise standard data;

[0094] In this embodiment, the original data of an enterprise includes text data corresponding to the information of the business location of the enterprise, the form of the enterprise's assets, the information of the enterprise's personnel, the information of the enterprise's business activities, the information of the enterprise's business qualifications, and the information of the enterprise's risks. These text data have not been classified and sorted, and the source is government public data, or provided by a third party and obtained by a web crawler. Among them, the government public data includes business registration information, tax information, and annual reports, etc. The third party provides detailed survey data collected including the financial status, personnel composition, and business activities of the enterprise. The web crawler includes enterprise-related information captured from channels such as enterprise websites, social media, and news reports on the Internet. Data cleaning refers to the process of re-examining and verifying data. Through data cleaning, the original data of the enterprise is processed to remove duplicate records, eliminate abnormal data, and correct erroneous data to ensure data consistency and improve data quality.

[0095] Step S102, inputting the enterprise standard data into a pre-trained classification model to obtain enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information;

[0096] In this embodiment, the pre-trained classification model can be a decision tree model trained by a large number of labeled sample data sets. The decision tree model is a classification and regression algorithm based on a tree structure. The decision tree consists of nodes (including root nodes, internal nodes and leaf nodes) and branches. The root node contains the entire sample data set, each internal node represents a test on a feature attribute, and each leaf node represents a category (for a classification tree) or a value (for a regression tree). At each node, the algorithm selects an optimal feature for splitting, divides the data into two or more subsets, and recursively repeats this process on these subsets. This process continues until a certain stopping criterion is met (such as samples in the subset all belong to the same category, or no more features are available). The classification training of the sample data set is completed through the above process.

[0097] Step S103, associating the enterprise business location information, the enterprise asset form information, the enterprise personnel information, the enterprise business activity information, the enterprise business qualification information, and the enterprise risk information with the corresponding enterprise to obtain the multi-dimensional information of the enterprise.

[0098] In this embodiment, the basic information of the enterprise is matched with the information identifiers corresponding to the enterprise's business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information. When the identification information identifier and the basic information of the enterprise are matched successfully, it means that the information corresponding to the identification information identifier corresponds to the enterprise, thereby completing the association between the enterprise's business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, enterprise risk information and the corresponding enterprise.

[0099] This embodiment obtains original enterprise data; inputs the original enterprise data into a pre-trained classification model to obtain enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information; associates the enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information with corresponding enterprises, thereby being able to obtain multi-dimensional enterprise information that is effectively classified based on the original enterprise data and associated with the enterprises, so as to facilitate the subsequent measurement of the value of enterprise information from multiple dimensions.

[0100] Continue to refer Figure 4In some optional implementations of this embodiment, step S20 includes the following steps:

[0101] Step S201, classifying the enterprise multidimensional information into variables to obtain enterprise multidimensional numerical variables and enterprise multidimensional categorical variables;

[0102] In this embodiment, variable classification is based on the variable type of enterprise multidimensional information, and the variable type includes numerical variables and categorical variables, wherein numerical variables are variables whose values ​​have computational significance, and are usually used to represent quantity, size, degree, etc., and categorical variables are variables whose values ​​have classification significance, and are mainly used to represent the attributes, type or state of things, etc. Variable classification is performed on enterprise multidimensional information based on the above numerical variables and categorical variables, thereby obtaining enterprise multidimensional numerical variables and enterprise multidimensional categorical variables. In this embodiment, the above enterprise multidimensional numerical variables include enterprise asset form information, and enterprise multidimensional categorical variables include enterprise business location information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information.

[0103] Step S202, normalizing the enterprise multidimensional numerical variables to obtain enterprise multidimensional standard numerical values;

[0104] In this embodiment, normalization is to convert variables into values ​​of a uniform scale. Since the scales of different numerical variables differ greatly, they need to be normalized to obtain values ​​that are convenient for subsequent calculations. Common normalization methods include minimum-maximum normalization (scaling values ​​to between 0 and 1) or standard normalization (using mean and standard deviation to convert values ​​to a standard normal distribution). In this embodiment, standard normalization can be used to normalize the enterprise's multidimensional numerical variables to obtain the enterprise's multidimensional standard numerical values.

[0105] Step S203, performing dummy variable processing on the enterprise multidimensional category variable to obtain enterprise multidimensional category information;

[0106] In this embodiment, categorical variables cannot usually be used directly for numerical calculations, so they need to be converted. Dummy variables (also called one-hot encoding) are a commonly used conversion method that converts each categorical variable into a set of new binary variables (also called dummy variables), where each dummy variable corresponds to a category. For example, if there is a categorical variable "financial indicators" whose values ​​include "sales", "profit", "cost" and "total assets", then after dummy variable processing, four new dummy variables will be obtained, indicating whether they belong to "sales", "profit", "cost" and "total assets" respectively. By using the categorical variables processed by dummy variables as enterprise multidimensional category information, the enterprise multidimensional category information can be used for feature extraction and model training together with the above-mentioned enterprise multidimensional standard values.

[0107] Step S204, associating the enterprise multi-dimensional standard value with the enterprise multi-dimensional category information to obtain enterprise multi-dimensional valid information;

[0108] In this embodiment, information association can be formed by combining two types of data, enterprise multidimensional standard values ​​and enterprise multidimensional category information, into a bar graph set, where each bar graph represents enterprise multidimensional category information and the height of each bar graph represents the enterprise multidimensional standard value.

[0109] Step S205, inputting the enterprise's multi-dimensional valid information into the pre-trained feature extraction model to obtain the enterprise's multi-dimensional features.

[0110] In this embodiment, the pre-trained feature extraction model can adopt the above-mentioned decision tree model, or can adopt a machine learning model capable of classification and feature extraction, such as a neural network model, a clustering algorithm model, etc. In this embodiment, the multi-dimensional features of the enterprise include geographical location features, asset structure features, asset value features, asset operation efficiency features, personnel scale features, personnel structure features, personnel efficiency features, business scope features, market performance features, sales strategy features, qualification level features, qualification coverage features, compliance features, credit risk features, operational risk features, etc.

[0111] This embodiment obtains enterprise multidimensional numerical variables and enterprise multidimensional categorical variables by classifying variables of the enterprise multidimensional information; normalizes the enterprise multidimensional numerical variables to obtain enterprise multidimensional standard numerical values; performs dummy variable processing on the enterprise multidimensional categorical variables to obtain enterprise multidimensional categorical information; associates the enterprise multidimensional standard numerical values ​​with the enterprise multidimensional categorical information to obtain enterprise multidimensional valid information; inputs the enterprise multidimensional valid information into the pre-trained feature extraction model, thereby obtaining enterprise multidimensional valid information corresponding to different types of variables, and obtaining valuable enterprise multidimensional features based on the enterprise multidimensional valid information, so as to facilitate subsequent scoring processing of the features.

[0112] Continue to refer Figure 5 In some optional implementations of this embodiment, step S201 includes the following steps:

[0113] Step S2011, obtaining multi-dimensional information of the enterprise;

[0114] In this embodiment, the enterprise multi-dimensional information is stored in a database, and the enterprise multi-dimensional information can be acquired by generating an information query request corresponding to the enterprise multi-dimensional information and extracting information from the database according to the information query request.

[0115] Step S2012, performing data cleaning and preprocessing on the enterprise multi-dimensional information to obtain standard enterprise multi-dimensional information;

[0116] In this embodiment, data cleaning and preprocessing of enterprise multidimensional information is an operation used to ensure data accuracy and consistency, and data cleaning and preprocessing include the following processing steps: missing value processing: checking missing values ​​in the data and deciding whether to delete records containing missing values, fill missing values ​​with default values, or estimate missing values ​​by some method (such as interpolation, machine learning prediction, etc.); outlier processing: identifying and processing outliers in the data, which may be caused by data entry errors, measurement errors, or other reasons. In this embodiment, the method of filling missing values ​​in enterprise multidimensional information is to process the missing values ​​with the average value to fill the missing values ​​in the enterprise multidimensional information. The outlier processing of enterprise multidimensional information is to identify and correct or delete outliers in the enterprise multidimensional information, which are commonly found in enterprise asset form information.

[0117] Step S2013, identifying the variable type corresponding to the standard enterprise multi-dimensional information, wherein the variable type includes a numerical variable and a categorical variable;

[0118] In this embodiment, standard enterprise multidimensional information can be identified through preset variable definitions. For example, the variable definition defines information that can be numerically calculated as a numerical variable, and defines information that cannot be numerically calculated as a categorical variable.

[0119] Step S2014, classifying the standard enterprise multidimensional information according to the numerical variables and the categorical variables to obtain the enterprise multidimensional numerical variables and the enterprise multidimensional categorical variables.

[0120] This embodiment obtains the enterprise multidimensional information; performs data cleaning and preprocessing on the enterprise multidimensional information to obtain standard enterprise multidimensional information; identifies the variable type corresponding to the standard enterprise multidimensional information, wherein the variable type includes numerical variables and categorical variables; and classifies the standard enterprise information according to the numerical variables and the categorical variables, thereby obtaining enterprise multidimensional numerical variables and enterprise multidimensional categorical variables classified according to the numerical types and categorical types corresponding to the information, to facilitate subsequent processing.

[0121] In some optional implementations of this embodiment, before step S20, the following steps are further included:

[0122] Acquire multidimensional information of sample enterprises, and input the multidimensional information of sample enterprises into a preset model to obtain multidimensional features of the sample enterprises;

[0123] In this embodiment, the multi-dimensional information of the sample enterprises includes the sample enterprise business location information, the sample enterprise asset form information, the sample enterprise personnel information, the sample enterprise business activity information, the sample enterprise business qualification information, and the sample enterprise risk information. The preset model is one of the above-mentioned decision tree models, neural network models, clustering algorithm models, etc., which can perform classification and feature extraction.

[0124] Calculate the multidimensional total score of the sample enterprises according to the multidimensional characteristics of the sample enterprises;

[0125] In this embodiment, the method of calculating the multidimensional total score of the sample enterprise by the multidimensional characteristics of the sample enterprise is the same as the method of calculating the multidimensional total score of the enterprise by the multidimensional characteristics of the above-mentioned enterprise. Different scoring standards and weights are set corresponding to the multidimensional characteristics of each sample enterprise. The multidimensional score of the sample enterprise is calculated by the multidimensional characteristics of the sample enterprise, and then the multidimensional score of the sample enterprise is weightedly calculated to obtain the multidimensional total score of the sample enterprise.

[0126] Determine whether the multi-dimensional total score of the sample enterprise is less than or equal to a preset threshold;

[0127] In this embodiment, if the multi-dimensional total score of the sample enterprise is greater than the preset threshold, the sample enterprise corresponding to the multi-dimensional total score of the sample enterprise is identified as a non-sample shell enterprise.

[0128] If the multi-dimensional total score of the sample enterprise is less than or equal to the preset threshold, the sample enterprise corresponding to the multi-dimensional total score of the sample enterprise is identified as a sample shell enterprise;

[0129] Obtaining historical shell company information, and judging whether the sample shell company is a real shell company based on the historical shell company information;

[0130] In this embodiment, if the sample shell company belongs to the real shell company, the currently preset model is used as the pre-trained feature extraction model.

[0131] If the sample shell enterprise does not belong to the real shell enterprise, the preset model is adjusted, and the multidimensional features of the new sample enterprise are re-acquired to calculate the multidimensional total score of the new sample enterprise, and the new sample shell enterprise is re-judged according to the multidimensional total score of the sample enterprise, and then the new sample shell enterprise is re-judged according to the historical shell enterprise information whether it belongs to the real shell enterprise, until the new sample shell enterprise belongs to the real shell enterprise;

[0132] In this embodiment, the above-mentioned step of adjusting the model until the sample shell companies belong to real shell companies is an iterative process of the model. In this embodiment, real shell companies can also be scored to obtain the total score of real shell companies. The multi-dimensional total score of the sample companies is iterated to a score range close to the total score of the real shell companies through the model, thereby completing the iterative process of the model. In the above-mentioned iterative process, a qualified threshold can be set, and the total score difference is obtained by calculating the difference between the multi-dimensional total score of the sample companies and the total score of the real shell companies, and then judging whether the total score difference is less than or equal to the qualified threshold. If the total score difference is less than or equal to the qualified threshold, the model iteration process is completed. If the total score difference is greater than the qualified threshold, the model iteration process is continued until the total score difference is less than or equal to the qualified threshold.

[0133] In this embodiment, the qualified threshold can be set according to the following method: First, according to the business background and the definition of shell companies, a reasonable qualified threshold range is preliminarily set. This range should take into account business tolerance, natural fluctuations in historical data, and uncertainty in model prediction. Then set a preliminary threshold, use the preliminary threshold to iterate the model, and dynamically adjust the qualified threshold during the model iteration process according to the difference between the multi-dimensional total score of the sample enterprise and the total score of the real shell enterprise after each iteration. For example, if the total score difference is much higher than the preliminarily set threshold in multiple consecutive iterations, and the model converges slowly, consider appropriately relaxing the threshold to promote the model to enter a stable state faster. Convergence judgment: monitor the trend of the total score difference during the iteration process. If the total score difference tends to be stable in multiple consecutive iterations and fluctuates within a small range, it can be regarded as the model is close to the optimal state, and the threshold at this time can be regarded as a more appropriate threshold. Using the cross-validation method, the data set is divided into a training set and a validation set, the model is iterated on the training set and the threshold is adjusted, and then the adjusted threshold is verified on the validation set to effectively distinguish between real shell companies and non-shell companies. In combination with actual business needs, set a series of evaluation indicators (such as accuracy, recall, F1 score, etc.), evaluate the performance of the model under different thresholds, and select the threshold that best meets business needs. Regularly review the effectiveness of the model and thresholds, and make necessary adjustments based on emerging business scenarios and data changes. Establish an abnormal monitoring mechanism, and intervene in analysis and adjust the model or threshold in a timely manner when abnormal fluctuations occur in the model output or total score difference.

[0134] After the new sample shell company belongs to the real shell company, the preset model is used as the pre-trained feature extraction model.

[0135] This embodiment obtains multidimensional information of sample enterprises and inputs the multidimensional information of sample enterprises into a preset model to obtain multidimensional features of sample enterprises; calculates the multidimensional total score of sample enterprises according to the multidimensional features of sample enterprises; determines whether the multidimensional total score of sample enterprises is less than or equal to a preset threshold; if the multidimensional total score of sample enterprises is less than or equal to the preset threshold, the sample enterprise corresponding to the multidimensional total score of sample enterprises is identified as a sample shell enterprise; obtains historical shell enterprise information, and determines whether the sample shell enterprise belongs to a real shell enterprise according to the historical shell enterprise information; if the sample shell enterprise does not belong to the real shell enterprise, the preset model is adjusted until the new sample shell enterprise belongs to the real shell enterprise; after the new sample shell enterprise belongs to the real shell enterprise, the preset model is used as the pre-trained feature extraction model. Thus, the model can be trained and iteratively optimized according to the multidimensional information of sample enterprises to obtain a pre-trained model that can effectively extract features.

[0136] Continue to refer Figure 6 In some optional implementations of this embodiment, step S30 includes the following steps:

[0137] Step S301, obtaining the enterprise category identifier corresponding to the enterprise multi-dimensional feature;

[0138] In this embodiment, the enterprise category identification is identification information for classifying enterprises corresponding to the multi-dimensional characteristics of the enterprise. Enterprises can be classified according to different business scopes, such as manufacturing, service, information technology, finance, etc., and can also be further subdivided according to size, such as large, medium, small or micro enterprises, and classified according to business models, such as wholesale and retail, production and manufacturing, research and development design, technical services, platform operations, etc.

[0139] Step S302, extracting the preset scoring criteria from the database according to the enterprise category identifier;

[0140] In this embodiment, the preset scoring standard is acquired in the above-mentioned acquisition manner and stored in the database, and the preset scoring standard can be extracted from the database according to the enterprise category identifier.

[0141] Step S302, mapping the enterprise multidimensional features to the preset scoring criteria to obtain the enterprise multidimensional score.

[0142] In this embodiment, by matching and corresponding the enterprise multidimensional features with the preset scoring criteria, each feature is associated with a corresponding indicator in the scoring criteria to complete the mapping of the enterprise multidimensional features with the preset scoring criteria.

[0143] This embodiment extracts the preset scoring criteria from the database and maps the enterprise multi-dimensional features to the preset scoring criteria, thereby effectively obtaining accurate enterprise multi-dimensional scoring.

[0144] Continue to refer Figure 7 In some optional implementations of this embodiment, step S40 includes the following steps:

[0145] Step S401, obtaining the score weight corresponding to the multi-dimensional score of the enterprise;

[0146] In this embodiment, the above-mentioned scoring weights can be obtained by inputting historical enterprise data into the model for analysis. The historical enterprise data includes historical shell companies and historical non-shell companies. The model can adopt a logistic regression or a neural network model to set the scoring weights corresponding to the multidimensional features of each enterprise by learning the characteristics of historical shell companies, and use the scoring weights corresponding to the multidimensional features of the enterprise as the scoring weights.

[0147] Step S402, weighting the enterprise business location score, the enterprise asset form score, the enterprise personnel score, the enterprise business activity score, the enterprise business qualification score, and the enterprise risk score according to the score weight;

[0148] In this embodiment, for each scoring item (business location score, enterprise asset form score, enterprise personnel score, enterprise business activity score, enterprise business qualification score, enterprise risk score), its score value is multiplied by the corresponding score weight to obtain a weighted score. For example, if the business location score of the enterprise is 80 points and its corresponding score weight is 0.3, the weighted score is 80*0.3=24 points.

[0149] Step S403, summing up the weighted scores of the enterprise's business location, the enterprise's asset form, the enterprise's personnel, the enterprise's business activity, the enterprise's business qualification, and the enterprise's risk score to obtain the enterprise's multi-dimensional total score.

[0150] In this embodiment, all weighted scoring items are added together to obtain the enterprise's multidimensional total score. For example, if the weighted enterprise business location score is 12 points, the enterprise asset form score is 15 points, the enterprise personnel score is 10 points, the enterprise business activity score is 15 points, the enterprise business qualification score is 8 points, and the enterprise risk score is 10 points, then the enterprise's multidimensional total score is 12+15+10+15+8+10=70 points.

[0151] This embodiment obtains the scoring weights corresponding to the multi-dimensional scoring of the enterprise; weights the enterprise business location score, the enterprise asset type score, the enterprise personnel score, the enterprise business activity score, the enterprise business qualification score, and the enterprise risk score according to the scoring weights; and sums the weighted enterprise business location score, the enterprise asset type score, the enterprise personnel score, the enterprise business activity score, the enterprise business qualification score, and the enterprise risk score, thereby effectively obtaining the multi-dimensional total score of the enterprise for comprehensive multi-dimensional data evaluation, so as to provide a valuable basis for subsequent enterprise identification and judgment.

[0152] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions 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-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0153] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0154] Further references Figure 8 , as a response to the above Figure 1 In order to realize the method shown in the figure, the present application provides an embodiment of a corporate identification device, and the device embodiment is Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0155] like Figure 8 As shown, the enterprise identification device 800 described in this embodiment includes: an information acquisition module 801, a feature acquisition module 802, a feature scoring module 803, a scoring calculation module 804, a total score determination module 805, a first result identification module 806, and a second result identification module 807. Among them:

[0156] Information acquisition module 801, used to acquire multi-dimensional information of the enterprise;

[0157] A feature acquisition module 802 is used to input the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain enterprise multi-dimensional features;

[0158] The feature scoring module 803 is used to score the enterprise multidimensional features according to a preset scoring standard to obtain the enterprise multidimensional score;

[0159] The score calculation module 804 is used to perform weighted calculation on the enterprise multi-dimensional score to obtain the enterprise multi-dimensional total score;

[0160] Total score determination module 805, used to determine whether the enterprise multi-dimensional total score is less than or equal to a preset threshold;

[0161] The first result identification module 806 is used to identify the enterprise corresponding to the multi-dimensional total score of the enterprise as a shell enterprise if the multi-dimensional total score of the enterprise is less than or equal to the preset threshold.

[0162] The second result identification module 807 is used to identify the enterprise corresponding to the multi-dimensional total score of the enterprise as a non-shell enterprise if the multi-dimensional total score of the enterprise is greater than the preset threshold.

[0163] This embodiment can obtain enterprise multidimensional information by adopting the above-mentioned enterprise identification device; input the enterprise multidimensional information into the pre-trained feature extraction model for feature extraction to obtain enterprise multidimensional features; score the enterprise multidimensional features according to the preset scoring standard to obtain the enterprise multidimensional score; perform weighted calculation on the enterprise multidimensional score to obtain the enterprise multidimensional total score; determine whether the enterprise multidimensional total score is less than or equal to the preset threshold; if the enterprise multidimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a shell enterprise; if the enterprise multidimensional total score is greater than the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a non-shell enterprise. This embodiment can extract features based on multidimensional data related to the enterprise through the above method, and perform weighted scoring on each feature, so as to obtain the enterprise multidimensional total score reflecting the enterprise risk status, and effectively identify shell enterprises based on the enterprise multidimensional total score, thereby completing the accurate identification of shell enterprises, reducing the risks in the process of carrying out financial business, and effectively ensuring the normal development of financial business.

[0164] To solve the above technical problems, the present application also provides a computer device. Fig. 9 , Fig. 9 This is a basic structural block diagram of the computer device in this embodiment.

[0165] The computer device 9 includes a memory 91, a processor 92, and a network interface 93 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 9 with components 91-93, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0166] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0167] The memory 91 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 91 can be an internal storage unit of the computer device 9, such as a hard disk or memory of the computer device 9. In other embodiments, the memory 91 can also be an external storage device of the computer device 9, such as a plug-in hard disk equipped on the computer device 9, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 91 can also include both the internal storage unit of the computer device 9 and its external storage device. In this embodiment, the memory 91 is generally used to store the operating system and various application software installed on the computer device 9, such as computer-readable instructions of the enterprise identification method, etc. In addition, the memory 91 can also be used to temporarily store various types of data that have been output or are to be output.

[0168] The processor 92 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 92 is generally used to control the overall operation of the computer device 9. In this embodiment, the processor 92 is used to run computer-readable instructions stored in the memory 91 or process data, such as computer-readable instructions for running the enterprise identification method.

[0169] The network interface 93 may include a wireless network interface or a wired network interface. The network interface 93 is generally used to establish a communication connection between the computer device 9 and other electronic devices.

[0170] By adopting the above-mentioned computer equipment, this embodiment can obtain enterprise multidimensional information; input the enterprise multidimensional information into the pre-trained feature extraction model for feature extraction to obtain enterprise multidimensional features; score the enterprise multidimensional features according to the preset scoring standard to obtain the enterprise multidimensional score; perform weighted calculation on the enterprise multidimensional score to obtain the enterprise multidimensional total score; determine whether the enterprise multidimensional total score is less than or equal to the preset threshold; if the enterprise multidimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a shell enterprise; if the enterprise multidimensional total score is greater than the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a non-shell enterprise. Through the above-mentioned method, this embodiment can perform feature extraction based on multidimensional data related to the enterprise, and perform weighted scoring on each feature, so as to obtain the enterprise multidimensional total score reflecting the enterprise risk status, and effectively identify shell enterprises based on the enterprise multidimensional total score, thereby completing the accurate identification of shell enterprises, reducing the risk in the process of carrying out financial business, and effectively ensuring the normal development of financial business.

[0171] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the enterprise identification method as described above.

[0172] This embodiment can obtain enterprise multidimensional information by adopting the above-mentioned computer-readable storage medium; input the enterprise multidimensional information into the pre-trained feature extraction model for feature extraction to obtain enterprise multidimensional features; score the enterprise multidimensional features according to the preset scoring criteria to obtain the enterprise multidimensional score; perform weighted calculation on the enterprise multidimensional score to obtain the enterprise multidimensional total score; determine whether the enterprise multidimensional total score is less than or equal to the preset threshold; if the enterprise multidimensional total score is less than or equal to the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a shell enterprise; if the enterprise multidimensional total score is greater than the preset threshold, the enterprise corresponding to the enterprise multidimensional total score is identified as a non-shell enterprise. This embodiment can extract features based on multidimensional data related to the enterprise through the above method, and perform weighted scoring on each feature, so as to obtain the enterprise multidimensional total score reflecting the enterprise risk status, and effectively identify shell enterprises based on the enterprise multidimensional total score, thereby completing the accurate identification of shell enterprises, reducing the risks in the process of carrying out financial business, and effectively ensuring the normal development of financial business.

[0173] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0174] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for enterprise identification, characterized in that: The steps include: Obtain multi-dimensional information of enterprises; Inputting the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain enterprise multi-dimensional features; Scoring the multidimensional characteristics of the enterprise according to a preset scoring standard to obtain a multidimensional score of the enterprise; Performing weighted calculation on the enterprise's multidimensional scores to obtain the enterprise's multidimensional total score; Determine whether the enterprise's multi-dimensional total score is less than or equal to a preset threshold; If the multi-dimensional total score of the enterprise is less than or equal to the preset threshold, the enterprise corresponding to the multi-dimensional total score of the enterprise is identified as a shell enterprise.

2. The enterprise identification method according to claim 1, characterized in that: The step of obtaining multi-dimensional information of an enterprise specifically includes: Acquire original enterprise data, perform data cleaning on the original enterprise data, and obtain enterprise standard data; Input the enterprise standard data into the pre-trained classification model to obtain enterprise business location information, enterprise asset form information, enterprise personnel information, enterprise business activity information, enterprise business qualification information, and enterprise risk information; The enterprise business location information, the enterprise asset form information, the enterprise personnel information, the enterprise business activity information, the enterprise business qualification information, and the enterprise risk information are associated with the corresponding enterprise to obtain the multi-dimensional information of the enterprise.

3. The enterprise identification method according to claim 1, characterized in that: The step of inputting the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain the enterprise multi-dimensional features specifically includes: Classifying the enterprise multidimensional information into variables to obtain enterprise multidimensional numerical variables and enterprise multidimensional categorical variables; Normalizing the enterprise's multidimensional numerical variables to obtain the enterprise's multidimensional standard numerical values; Performing dummy variable processing on the enterprise multidimensional category variables to obtain enterprise multidimensional category information; Associating the enterprise multi-dimensional standard value with the enterprise multi-dimensional category information to obtain enterprise multi-dimensional valid information; The enterprise's multi-dimensional valid information is input into the pre-trained feature extraction model to obtain the enterprise's multi-dimensional features.

4. The enterprise identification method according to claim 3, characterized in that: The step of classifying the enterprise multidimensional information into variables to obtain enterprise multidimensional numerical variables and enterprise multidimensional categorical variables specifically includes: Acquiring multi-dimensional information of the enterprise; Performing data cleaning and preprocessing on the enterprise multidimensional information to obtain standard enterprise multidimensional information; Identify the variable type corresponding to the standard enterprise multi-dimensional information, wherein the variable type includes a numerical variable and a categorical variable; The standard enterprise multidimensional information is classified according to the numerical variables and the categorical variables to obtain the enterprise multidimensional numerical variables and the enterprise multidimensional categorical variables.

5. The enterprise identification method according to claim 1, characterized in that: Before the step of inputting the enterprise multi-dimensional information into the pre-trained feature extraction model to extract features and obtain the enterprise multi-dimensional features, the following steps are also included: Acquire multidimensional information of sample enterprises, and input the multidimensional information of sample enterprises into a preset model to obtain multidimensional features of the sample enterprises; Calculate the multidimensional total score of the sample enterprises according to the multidimensional characteristics of the sample enterprises; Determine whether the multi-dimensional total score of the sample enterprise is less than or equal to a preset threshold; If the multi-dimensional total score of the sample enterprise is less than or equal to the preset threshold, the sample enterprise corresponding to the multi-dimensional total score of the sample enterprise is identified as a sample shell enterprise; Obtaining historical shell company information, and judging whether the sample shell company is a real shell company based on the historical shell company information; If the sample shell enterprise does not belong to the real shell enterprise, the preset model is adjusted, and the multidimensional features of the new sample enterprise are re-acquired to calculate the multidimensional total score of the new sample enterprise, and the new sample shell enterprise is re-judged according to the multidimensional total score of the sample enterprise, and then the new sample shell enterprise is re-judged according to the historical shell enterprise information whether it belongs to the real shell enterprise, until the new sample shell enterprise belongs to the real shell enterprise; After the new sample shell company belongs to the real shell company, the preset model is used as the pre-trained feature extraction model.

6. The enterprise identification method according to claim 1, characterized in that: The step of scoring the enterprise multidimensional characteristics according to the preset scoring criteria to obtain the enterprise multidimensional score specifically includes: Obtain the enterprise category identifier corresponding to the enterprise's multi-dimensional features; Extracting the preset scoring criteria from a database according to the enterprise category identifier; The multi-dimensional characteristics of the enterprise are mapped to the preset scoring criteria to obtain the multi-dimensional score of the enterprise.

7. The enterprise identification method according to claim 1, characterized in that: The enterprise multi-dimensional score includes the enterprise business location score, enterprise asset form score, enterprise personnel score, enterprise business activity score, enterprise business qualification score, and enterprise risk score. The step of weighted calculation of the enterprise multi-dimensional score to obtain the enterprise multi-dimensional total score specifically includes: Obtain the score weight corresponding to the multi-dimensional score of the enterprise; Weighting the enterprise business location score, the enterprise asset form score, the enterprise personnel score, the enterprise business activity score, the enterprise business qualification score, and the enterprise risk score according to the score weight; The weighted scores of the enterprise's business location, asset form, personnel, business activities, business qualifications and risk are summed to obtain the enterprise's multi-dimensional total score.

8. A corporate identification device, characterized in that: include: Information acquisition module, used to obtain multi-dimensional information of enterprises; A feature acquisition module is used to input the enterprise multi-dimensional information into a pre-trained feature extraction model to extract features and obtain enterprise multi-dimensional features; A feature scoring module is used to score the enterprise's multidimensional features according to a preset scoring standard to obtain a multidimensional score of the enterprise; A scoring calculation module, used to perform weighted calculation on the enterprise's multi-dimensional scores to obtain the enterprise's multi-dimensional total score; A total score judgment module is used to judge whether the multi-dimensional total score of the enterprise is less than or equal to a preset threshold; The result identification module is used to identify the enterprise corresponding to the multi-dimensional total score of the enterprise as a shell enterprise if the multi-dimensional total score of the enterprise is less than or equal to the preset threshold.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the enterprise identification method according to any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the enterprise identification method according to any one of claims 1 to 7 are implemented.

Images