Company credit risk assessment method and system, computer equipment and storage medium

Abstract

A company credit risk assessment method comprises the following steps that company information of a target company is acquired, and the company information comprises stockholder information, area position information, business scope information and business condition information; obtaining a plurality of similar companies similar to the target company based on the regional position information and the business scope information, and generating a regional adjustment factor based on the number of the similar companies; screening the similar companies based on shareholder information to obtain substantially associated companies, and acquiring operation condition information of the substantially associated companies; performing basic prediction on the target company based on the financial condition information by using a pre-trained risk prediction model to obtain a basic risk degree, and performing basic prediction on the substantially associated company to obtain an associated adjustment factor; and adjusting the basic risk degree by using the regional adjustment factor and the associated adjustment factor to obtain a final risk degree. The reliability of the obtained final risk degree is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of commercial credit assessment, and in particular to a method, system, computer equipment and storage medium for assessing corporate credit risk. Background Art

[0002] Credit is the respect and protection of the legitimate rights and interests of traders in the market economy, and it is also the basic line of defense to avoid transaction risks under information asymmetry. It is of great significance for enterprises to have a good financial situation through a perfect credit management system and credit risk monitoring methods. If a good judgment can be made on the credit of an enterprise in a commercial transaction scenario, then the success of commercial transactions and the healthy and stable development of various enterprises can be guaranteed to a considerable extent.

[0003] In the prior art, when assessing the credit risk of an enterprise, it mainly relies on the relevant operating data of the target company and uses a specific credit assessment model to assess the risk level. However, the accuracy of such methods is limited. The main reason is that it is very difficult to obtain relevant operating data, and it is difficult to obtain accurate results with one-sided data. Moreover, relying solely on relevant operating data for assessment is itself very one-sided. If a company's relevant operating data performs well, but the industry it operates has no advantages in the region where it is located, or it cannot obtain excellent resources in the region where it is located, then if there are market fluctuations, the company is also prone to passive operating problems. That is, when a company cannot rely on the resources of the local advantageous industry, its credit risk will be higher. The prior art method of credit risk assessment that relies solely on relevant operating data cannot take this factor into account. Summary of the invention

[0004] In order to address the deficiencies in the prior art, the present invention provides a corporate credit risk assessment method, system, computer equipment and storage medium, which jointly assess the credit risk of a target company based on the target company's operating status information and the industrial status of the region where it is located, and the resulting final risk level is more reliable.

[0005] In order to achieve the above object, the specific scheme adopted by the present invention is:

[0006] A corporate credit risk assessment method comprises the following steps:

[0007] Obtaining the target company's company information, including shareholder information, regional location information, business scope information, and business status information;

[0008] Acquire a plurality of similar companies similar to the target company based on the regional location information and the business scope information, and generate a regional adjustment factor based on the number of the similar companies;

[0009] Screening the similar companies based on the shareholder information to obtain substantially related companies, and acquiring the operating status information of the substantially related companies;

[0010] Using a pre-trained risk prediction model to perform a basic prediction on the target company based on the financial status information to obtain a basic risk level, and performing a basic prediction on the substantially associated company to obtain an associated adjustment factor;

[0011] The basic risk level is adjusted using the regional adjustment factor and the associated adjustment factor to obtain a final risk level.

[0012] Preferably, a specific method for acquiring a plurality of similar companies similar to the target company based on the regional location information and the business scope information includes:

[0013] Acquire multiple candidate companies belonging to the same region as the target company based on the regional location information;

[0014] Performing a first screening of the candidate companies based on the business scope information to obtain a plurality of first-screened companies;

[0015] The first-screened companies are screened a second time based on their registered scale to obtain a plurality of similar companies that are similar to the target company.

[0016] Preferably, the method for first screening the candidate companies based on the business scope information includes:

[0017] Calculate the similarity between the business scope information of the candidate companies and the business scope information of the target company one by one to obtain a similarity value;

[0018] If the similarity value corresponding to the candidate company reaches a preset similarity threshold, the candidate company is screened as the first-screened company.

[0019] Preferably, the method of conducting a second screening of the first screened companies based on the registered scale of the first screened companies comprises:

[0020] Obtaining the registered scale of the first screening company and the target company, wherein the registered scale includes registered capital and social security data;

[0021] Calculate the difference between the registered scale of the first-screened companies and the registered scale of the target companies one by one to obtain the scale difference;

[0022] If the size difference is lower than the preset difference threshold, the first screened company will be screened as the similar company.

[0023] Preferably, the method for generating a regional adjustment factor based on the number of similar companies includes:

[0024] Determine the existence duration of the similar companies, group the similar companies based on the existence duration, and generate multiple weights, where a group of the similar companies corresponds to one weight;

[0025] When the number of a group of similar companies reaches a preset number threshold, a benchmark factor is generated;

[0026] Optimizing the benchmark factor using the weight;

[0027] The regional adjustment factor is generated based on the optimized reference factor.

[0028] Preferably, the regional adjustment factor is used to reduce the basic risk level, and the associated adjustment factor is used to increase the basic risk level.

[0029] Preferably, the risk prediction model adopts a BP neural network model, and the method for training the risk prediction model includes:

[0030] Collecting a data set, the data set includes a plurality of company risk data, and dividing the data set into a training set, a validation set, and a test set;

[0031] Using the training set to train the basic BP neural network model, and using the verification set to evaluate the training status during the training process;

[0032] When the number of training iterations reaches the preset maximum number, the training is terminated, and the trained basic BP neural network is verified using the test set. If the verification passes, the risk prediction model is output.

[0033] A corporate credit risk assessment device, comprising:

[0034] An information acquisition module, used to acquire the company information of the target company;

[0035] A company screening module, used for acquiring a plurality of similar companies similar to the target company based on the regional location information and the business scope information, and screening the similar companies based on the shareholder information to obtain substantially related companies;

[0036] A data processing module, for generating a regional adjustment factor based on the number of similar companies;

[0037] The risk prediction module is used to use a pre-trained risk prediction model to make a basic prediction of the target company based on the financial status information to obtain a basic risk level, to make a basic prediction of the substantially associated company to obtain an associated adjustment factor, and to adjust the basic risk level using the regional adjustment factor and the associated adjustment factor to obtain a final risk level.

[0038] Computer equipment, including:

[0039] a memory storing a computer program;

[0040] The processor is used to read the computer program from the memory and execute it to implement the above-mentioned corporate credit risk assessment method.

[0041] The storage medium is used to store a computer program, and when the computer program is executed, the above-mentioned corporate credit risk assessment method is implemented.

[0042] The present invention first screens out similar companies in the region where the target company is located that are similar to the target company, then judges the industrial status of the region where the target company is located based on the similar companies, and finally evaluates the credit risk of the target company based on the target company's operating status information and the industrial status of the region, and the reliability of the final risk level obtained is stronger. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 is a flow chart of the company credit risk assessment method of the present invention;

[0045] Figure 2 It is a structural block diagram of a corporate credit risk assessment device. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1 , a corporate credit risk assessment method, including S1 to S5.

[0048] S1. Obtain the company information of the target company, which includes shareholder information, regional location information, business scope information and business status information. In the company information, the shareholder information is used to screen the substantially related companies with the same shareholders as the target company, and then the credit risk of the target company can be judged based on the credit record of the substantially related companies. The regional location information and business scope information are used to screen similar companies located in the same region as the target company and with similar business scopes. When there are a large number of companies with similar business scopes in a region, it means that the region has certain industrial advantages in the industries corresponding to these similar business scopes. These industrial advantages may be manifested in many upstream and downstream supporting enterprises, low raw material costs, low labor costs, or developed logistics industries. When a company belongs to a local advantageous industry, it will have more advantages than companies of the same level in other regions, and the risk of falling into business difficulties is also smaller. Therefore, the present invention uses regional location information and business scope information to screen similar companies similar to the target company, and assists in judging the credit risk of the target company based on the industrial situation in the region where the target company is located. The business status information is used to reflect the current business status of the target company. If the business status is good, the corresponding credit risk is also low. Therefore, the business status information is the main basis for judging the credit risk of the target company. The specific processing and application methods of company information are as S2 to S5.

[0049] S2. Based on the regional location information and business scope information, multiple similar companies similar to the target company are obtained, and a regional adjustment factor is generated based on the number of similar companies. When there are many similar companies in the region where the target company is located, it means that the target company may be operating a dominant industry in the region. The more similar companies there are, the larger the scale of the dominant industry is. The generated regional adjustment factor is used to reduce the credit risk of the target company.

[0050] The specific method of acquiring multiple similar companies similar to the target company based on the regional location information and the business scope information includes S21 to S23.

[0051] S21. Based on the regional location information, multiple candidate companies belonging to the same region as the target company are obtained. In S21, based on the registration place announced in the National Enterprise Credit Information Publicity System, the city-level administrative region can be used as the basis, and all companies registered in the same city as the target company can be used as candidate companies.

[0052] S22. Perform a first screening of candidate companies based on the business scope information to obtain multiple first-screened companies.

[0053] The method for first screening candidate companies based on business scope information includes S221 to S222.

[0054] S221. Calculate the similarity of the business scope information of the candidate companies and the business scope information of the target company one by one to obtain a similarity value. Because many companies do not have a single business scope, but have multiple different business scopes, it is necessary to judge the similarity of the business scopes of the target company and the candidate companies, that is, to calculate the similarity. Specifically, first screen out all the same business scopes of the target company and the candidate companies, then calculate the ratio of the same business scope to all the business scopes of the target company, and use the ratio as the similarity value.

[0055] S222. If the similarity value corresponding to the candidate company reaches the preset similarity threshold, the candidate company is screened as a first screening company. The higher the similarity value, the more common business scopes the candidate company has with the target company, which means that the target company and the candidate company are more likely to operate the same type of industry. When the similarity value exceeds the preset similarity threshold, it means that most of the business scopes of the target company are the same as those of the candidate company, and the candidate company is sufficiently similar to the target company. In the present invention, the similarity threshold can be set to 65-80%.

[0056] S23. Perform a second screening of the first-screened companies based on their registered scale, and obtain multiple similar companies that are similar to the target company. After screening out the candidate companies, there may be a situation where the target company and the first-screened company have the same business scope as the first-screened company, which accounts for too low a proportion of the first-screened company's business scope. The main reason for this is that the first-screened company is larger than the target company, resulting in the first-screened company also operating the same industry as the target company, but also operating many other industries at the same time. Under this scale difference, the first-screened company is still quite different from the target company, so a second screening is required to screen out similar companies that are truly similar to the target company.

[0057] The method of conducting a second screening of the first-screened companies based on the registered scale of the first-screened companies includes S231 to S233.

[0058] S231. Obtain the registered scale of the screening company and the target company, which includes the registered capital and social security data.

[0059] S232. Calculate the difference between the registered scale of the first screening company and the registered scale of the target company one by one to obtain the scale difference.

[0060] S233. If the size difference is lower than the preset difference threshold, the first-screened companies are screened as similar companies.

[0061] In S23, it is considered that the registered capital may not truly reflect the actual scale of the company after the company has been in operation for many years, and the social security data directly reflects the number of insured persons in the company. The more insured persons there are, the more employees the company has, and the more it can reflect the company's actual operating scale. Therefore, when calculating the difference, the registered capital and social security data are calculated separately, among which the social security data uses the number of insured persons, and the calculated difference value is still expressed in the form of a ratio. For example, if the registered capital of a screened company is n1 and the registered capital of the target company is n2, then (n1-n2) / n2 is calculated first, and the difference value of the registered capital is obtained after the absolute value is obtained.

[0062] After calculating the two difference values, different weights are assigned to the two difference values, among which the difference value of social security data has a higher weight. After taking the weighted average of the two difference values, the scale difference can be obtained.

[0063] If the size difference is small, it means that the target company and the once-screened companies have a small gap in registered size. By screening out these once-screened companies, similar companies can be obtained.

[0064] On the basis of obtaining similar companies, the method of generating regional adjustment factors based on the number of similar companies includes S24 to S27.

[0065] S24. Determine the duration of similar companies, group similar companies based on the duration, and generate multiple weights, with one weight corresponding to a group of similar companies. If there are many similar companies to the target company, but the duration is very short, these similar companies are all concentrated in a short period of time, and may lack real business activities. Although these similar companies operate the same industry as the target company, these industries may not be strong enough, that is, they may not be able to develop benignly. Although they can reach a certain scale in a short period of time, they are also likely to weaken rapidly. Therefore, it is necessary to consider the duration of similar companies to determine whether the region where the target company is located has real and reliable industrial advantages. Specifically, similar companies with a duration of 1 to 3 years can be grouped into one group, similar companies with a duration of 3 to 6 years can be grouped into one group, and similar companies with a duration of 6 to 10 years can be grouped into one group. The specific division method can be flexibly selected according to actual needs. The longer the duration of the group, the greater the corresponding weight.

[0066] S25. When the number of a group of similar companies reaches a preset number threshold, a benchmark factor is generated. When the number of a group of similar companies is small, it can be directly ignored to simplify the complexity of the calculation.

[0067] S26. Optimize the benchmark factors using weights.

[0068] S27. Generate regional adjustment factors based on the optimized benchmark factors. The more similar companies with longer duration are among similar companies, the more the regional adjustment factors generated by the optimized benchmark factors can reduce the risk level of the target company. Conversely, the more similar companies with shorter duration are among similar companies, the less the regional adjustment factors generated by the optimized benchmark factors can reduce the risk level of the target company.

[0069] S3. Screen similar companies based on shareholder information to obtain substantially related companies, and obtain operating information of substantially related companies. If there is a company with the same shareholders as the target company, and the company has serious credit risk or may have serious credit risk, then the target company is more likely to have credit risk, so it is necessary to evaluate the substantially related companies.

[0070] S4. Use the pre-trained risk prediction model to make a basic prediction of the target company based on the financial status information to obtain the basic risk level, and make a basic prediction of the actual associated company to obtain the associated adjustment factor. Because the company's operating status information involves a lot of data, including sales, profit, profit margin, market share and liabilities, etc., and the company also has many types of credit risks, including debt default, intellectual property disputes and ownership disputes, etc., it is difficult to intuitively determine the association between operating status information and credit risk. Therefore, the present invention adopts a neural network model and uses big data to predict the basic risk level. In the present invention, the risk prediction model adopts a BP neural network model, and the method for training the risk prediction model includes: collecting a data set, the data set includes a plurality of company risk data, and dividing the data set into a training set, a verification set and a test set; using the training set to train the basic BP neural network model, and using the verification set to evaluate the training status during the training process; when the number of training iterations reaches the preset maximum number, the training is terminated, and the trained basic BP neural network is verified using the test set, and if the verification passes, the risk prediction model is output. The input of the trained risk prediction model is the company's operating status information, and the output is the basic risk level. The specific training method of the basic BP neural network belongs to the mature existing technology and will not be described in detail here.

[0071] S5. Use the regional adjustment factor and the correlation adjustment factor to adjust the basic risk level to obtain the final risk level. The regional adjustment factor is used to reduce the basic risk level, and the correlation adjustment factor is used to increase the basic risk level.

[0072] In the present invention, the basic risk level and the final risk level can be expressed by specific numerical values, and accordingly, the final value range of the regional adjustment factor is (0, 1), and the final value range of the associated adjustment factor is greater than 1. The basic risk level and the final risk level can also be expressed in the form of levels, and accordingly, the regional adjustment factor is used to reduce the level, and the associated adjustment factor is used to increase the level.

[0073] See also Figure 2 The present invention further provides a corporate credit risk assessment device, including an information acquisition module, a company screening module, a data processing module and a risk prediction module.

[0074] The information acquisition module is used to obtain the company information of the target company.

[0075] The company screening module is used to obtain multiple similar companies similar to the target company based on regional location information and business scope information, and to screen similar companies based on shareholder information to obtain substantially related companies.

[0076] A data processing module for generating regional adjustment factors based on the number of similar companies.

[0077] The risk prediction module is used to use a pre-trained risk prediction model to make a basic prediction of the target company based on financial status information to obtain a basic risk level, to make a basic prediction of the substantially associated company to obtain an associated adjustment factor, and to adjust the basic risk level using a regional adjustment factor and an associated adjustment factor to obtain a final risk level.

[0078] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in the present disclosure, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0079] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0080] The present invention also provides a computer device, including a memory and a processor.

[0081] A memory storing a computer program.

[0082] The processor is used to read the computer program from the memory and execute it to implement the above-mentioned corporate credit risk assessment method.

[0083] Finally, the present invention also provides a storage medium for storing a computer program, which implements the above-mentioned corporate credit risk assessment method when executed.

[0084] Based on the storage medium, the technical solution of the present disclosure, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0085] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0086] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A corporate credit risk assessment method, characterized in that: The steps include: Obtaining the target company's company information, including shareholder information, regional location information, business scope information, and business status information; Acquire a plurality of similar companies similar to the target company based on the regional location information and the business scope information, and generate a regional adjustment factor based on the number of the similar companies; Screening the similar companies based on the shareholder information to obtain substantially related companies, and acquiring the operating status information of the substantially related companies; Using a pre-trained risk prediction model to perform a basic prediction on the target company based on the financial status information to obtain a basic risk level, and performing a basic prediction on the substantially associated company to obtain an associated adjustment factor; The basic risk level is adjusted using the regional adjustment factor and the associated adjustment factor to obtain a final risk level.

2. A corporate credit risk assessment method as claimed in claim 1, characterized in that: A specific method for acquiring a plurality of similar companies similar to a target company based on the regional location information and the business scope information includes: Acquire multiple candidate companies belonging to the same region as the target company based on the regional location information; Performing a first screening of the candidate companies based on the business scope information to obtain a plurality of first-screened companies; The first-screened companies are screened a second time based on their registered scale to obtain a plurality of similar companies that are similar to the target company.

3. A corporate credit risk assessment method as claimed in claim 2, characterized in that: The method for first screening the candidate companies based on the business scope information includes: Calculate the similarity between the business scope information of the candidate companies and the business scope information of the target company one by one to obtain a similarity value; If the similarity value corresponding to the candidate company reaches a preset similarity threshold, the candidate company is screened as the first-screened company.

4. A corporate credit risk assessment method as claimed in claim 2, characterized in that: The method of conducting a second screening of the first-screened companies based on the registered scale of the first-screened companies includes: Obtaining the registered scale of the first screening company and the target company, wherein the registered scale includes registered capital and social security data; calculating the difference between the registered scale of the first screening company and the registered scale of the target company one by one to obtain the scale difference; If the size difference is lower than the preset difference threshold, the first screened company will be screened as the similar company.

5. A corporate credit risk assessment method as claimed in claim 1, characterized in that: The method of generating a regional adjustment factor based on the number of similar companies includes: Determine the existence duration of the similar companies, group the similar companies based on the existence duration, and generate multiple weights, where a group of the similar companies corresponds to one weight; When the number of a group of similar companies reaches a preset number threshold, a benchmark factor is generated; Optimizing the benchmark factor using the weight; The regional adjustment factor is generated based on the optimized reference factor.

6. A corporate credit risk assessment method as claimed in claim 1, characterized in that: The regional adjustment factor is used to reduce the basic risk level, and the association adjustment factor is used to increase the basic risk level.

7. A corporate credit risk assessment method as claimed in claim 1, characterized in that: The risk prediction model adopts a BP neural network model, and the method for training the risk prediction model includes: Collecting a data set, the data set includes a plurality of company risk data, and dividing the data set into a training set, a validation set, and a test set; Using the training set to train the basic BP neural network model, and using the verification set to evaluate the training status during the training process; When the number of training iterations reaches the preset maximum number, the training is terminated, and the trained basic BP neural network is verified using the test set. If the verification passes, the risk prediction model is output.

8. A corporate credit risk assessment device, characterized in that: include: An information acquisition module, used to acquire the company information of the target company; A company screening module, used for acquiring a plurality of similar companies similar to the target company based on the regional location information and the business scope information, and screening the similar companies based on the shareholder information to obtain substantially related companies; A data processing module, for generating a regional adjustment factor based on the number of similar companies; The risk prediction module is used to use a pre-trained risk prediction model to make a basic prediction of the target company based on the financial status information to obtain a basic risk level, to make a basic prediction of the substantially associated company to obtain an associated adjustment factor, and to adjust the basic risk level using the regional adjustment factor and the associated adjustment factor to obtain a final risk level.

9. Computer device, characterized in that include: a memory storing a computer program; A processor is used to read the computer program from the memory and execute it to implement a corporate credit risk assessment method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that Used to store a computer program, which, when executed, implements a corporate credit risk assessment method as described in any one of claims 1 to 7.

Images