Credit rating method and device, computer equipment and storage medium

Abstract

The invention relates to a credit rating method and device, computer equipment and a storage medium. The method comprises the following steps: determining at least one piece of quantitative index data and at least one piece of qualitative index data associated with a default behavior of a target object; respectively determining respective quantitative rating data of the at least one quantitative index data and respective qualitative rating data of the at least one qualitative index data; determining a credit evaluation mode for the target object according to a target industry opening of the industry field to which the target object belongs, and determining comprehensive rating data of the target object based on the quantitative rating data and the qualitative rating data according to the credit rating mode; the target default probability of the target object is determined based on the comprehensive rating data, and the credit rating result of the target object is obtained based on the mapping relation between the target default probability and the credit rating opened for the target industry, so that the pertinence of the rating result can be enhanced, and the accurate and reliable credit rating result is obtained.

Description

Technical Field

[0001] This application relates to the field of credit risk assessment technology, and in particular to a credit rating method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] Credit rating, as a core means of measuring the default risk of market participants, plays a crucial role in scenarios such as financial market pricing, investment decisions, and risk management.

[0003] As the bond market continues to expand, the types of market participants are becoming increasingly diversified. Based on different exposures, they can include entities with different attributes such as local government financing vehicles (LGFVs), financial institutions, general industry enterprises, and cross-industry groups. Credit rating of each market participant can provide investors with an objective and fair risk assessment reference.

[0004] However, existing credit rating methods yield ratings with low accuracy. Summary of the Invention

[0005] Therefore, it is necessary to provide a credit rating method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the accuracy of rating in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a credit rating method, the method comprising:

[0007] Identify at least one quantitative indicator and at least one qualitative indicator that link the target entity to the breach of contract.

[0008] Determine the quantitative rating data for each of the at least one quantitative indicator data and the qualitative rating data for each of the at least one qualitative indicator data;

[0009] Based on the target industry exposure of the target object to which the target object belongs, determine the credit evaluation method for the target object, and in accordance with the credit rating method, determine the comprehensive rating data of the target object based on the quantitative rating data and the qualitative rating data;

[0010] Based on the comprehensive rating data, the target default probability of the target object is determined, and based on the mapping relationship between the target default probability and the credit rating for the target industry exposure, the credit rating result of the target object is obtained.

[0011] In one embodiment, determining the quantitative rating data for each of the at least one quantitative indicator data and the quantitative rating data for each of the at least one qualitative indicator data includes:

[0012] For each quantitative indicator data, the target object group to which the target object belongs in multiple associated object combinations is determined based on the targeted quantitative indicator data, and the quantitative rating data of the targeted quantitative indicator data is determined based on the targeted quantitative indicator data, the target object group, and each of the associated object groups; the associated object group is obtained by dividing multiple candidate objects that belong to the same business type as the target object into data;

[0013] For each qualitative indicator data, a target qualitative interval is determined among multiple associated qualitative intervals to which the target qualitative indicator data belongs, and qualitative rating data for the target qualitative indicator data is obtained based on the mapping relationship between the target qualitative interval and the qualitative data.

[0014] In one embodiment, determining the target object combination to which the target object belongs among multiple related object combinations based on the targeted quantitative indicator data includes:

[0015] Obtain quantitative indicator data for each of multiple candidate objects, including the target object and belonging to the same business type as the target object;

[0016] Based on the data distribution characteristics of each of the quantitative indicator data, each of the candidate objects is divided into multiple related object combinations;

[0017] The quantitative indicator data of the target object are matched with the quantitative indicator data range of each of the associated object combinations to obtain the target object combination to which the target object belongs.

[0018] In one embodiment, determining the quantitative rating data for the targeted quantitative indicator data based on the targeted quantitative indicator data, the target object combination, and each of the associated object combinations includes:

[0019] For each group of associated objects, the combined quantitative data of the group of associated objects is determined based on the quantitative index data of each candidate object included in the group of associated objects.

[0020] The combined default rate of the targeted associated object combination is determined based on the individual default rates of each candidate object included in the targeted associated object combination.

[0021] When the combined quantitative data and the combined default rate meet the trend change conditions, the quantitative rating data of the targeted quantitative indicator data is obtained based on the combined default rate.

[0022] In one embodiment, determining the target default probability of the target object based on the comprehensive rating data includes:

[0023] Determine the mapping relationship between the comprehensive rating data and the probability of default, as well as multiple mapping coefficients for the mapping relationship;

[0024] For each mapping coefficient, based on the comprehensive rating data, the mapping relationship, and the mapping coefficient, the candidate default probability of the target object is determined;

[0025] Based on the industry default probability of the target object's industry sector, the target default probability is determined from the candidate default probabilities.

[0026] In one embodiment, the method further includes:

[0027] Determine at least one of the following: credit change information, external association information, and historical rating deviation information of the target object;

[0028] The credit rating result is corrected based on at least one of the credit change information, the external correlation information, and the historical rating deviation information to obtain the corrected credit rating result.

[0029] Secondly, this application also provides a credit rating device, the device comprising:

[0030] The data determination module is used to determine at least one quantitative indicator data and at least one qualitative indicator data that are associated with the target object and the breach of contract.

[0031] The indicator rating module is used to determine the quantitative rating data of each of the at least one quantitative indicator data and the qualitative rating data of each of the at least one qualitative indicator data.

[0032] The comprehensive rating module is used to determine the credit rating method for the target object based on the target industry exposure of the target object's industry sector, and to determine the comprehensive rating data of the target object based on the quantitative rating data and the qualitative rating data according to the credit rating method.

[0033] The credit rating module is used to determine the target default probability of the target object based on the comprehensive rating data, and to obtain the credit rating result of the target object based on the mapping relationship between the target default probability and the credit rating for the target industry exposure.

[0034] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0035] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0036] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0037] The aforementioned credit rating methods, apparatus, computer equipment, computer-readable storage media, and computer program products determine at least one quantitative indicator and at least one qualitative indicator related to the default behavior of the target entity; determine the quantitative rating data for each of the at least one quantitative indicator and the qualitative rating data for each of the at least one qualitative indicator; determine the credit evaluation method for the target entity based on its industry exposure in the industry sector to which it belongs, and determine the comprehensive rating data for the target entity based on the quantitative and qualitative rating data according to the credit rating method; determine the target default probability of the target entity based on the comprehensive rating data, and obtain the credit rating result of the target entity based on the mapping relationship between the target default probability and the credit rating for the target industry exposure. By identifying both quantitative and qualitative indicator data, and separately determining the quantitative rating data for the quantitative indicators and the qualitative rating data for the qualitative indicators, and then processing the quantitative and qualitative rating data in conjunction with the target entity's industry exposure, a comprehensive rating data for the target entity is obtained. Further, based on this comprehensive rating data, the target default probability of the target entity is determined, leading to the target entity's credit rating result. This approach comprehensively considers default-related factors of the target entity and conducts ratings within the corresponding target industry exposure, enhancing the relevance of the rating results and resulting in accurate and reliable credit rating outcomes. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a diagram illustrating the application environment of a credit rating method in one embodiment;

[0040] Figure 2 This is a flowchart illustrating a credit rating method in one embodiment;

[0041] Figure 3 This is a flowchart illustrating the process of determining quantitative rating data in one embodiment;

[0042] Figure 4 This is a flowchart illustrating a credit rating method in an application example.

[0043] Figure 5 This is a structural block diagram of a credit rating device in one embodiment;

[0044] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0047] In related technologies, credit rating methods based on financial data and information detected by each unit can be constructed using statistics or deep learning, such as logistic regression models, neural network models, Bayesian classification models, gradient boosting decision tree models, and random forest classification models. A trained model can be obtained by training the training dataset and then used to predict the credit rating results of the target entity.

[0048] However, most of these methods are based on financial data to construct statistical and machine learning methods to calculate the credit rating of bond issuers. In practice, because bond issuers with different exposures have different characteristics and sample sizes, the rating results obtained by using a single method have poor interpretability and low accuracy.

[0049] The credit rating method provided in this application embodiment can be applied to, for example, Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed on a cloud or other network server. When a user initiates a credit rating request through terminal 102, server 104 can determine at least one quantitative indicator and at least one qualitative indicator related to the default behavior of the target entity based on the credit rating request, and determine the quantitative rating data for each of the at least one quantitative indicator and the qualitative rating data for each of the at least one qualitative indicator. Server 104 determines the credit evaluation method for the target entity based on its industry exposure within its industry sector, and determines the comprehensive rating data for the target entity based on the quantitative and qualitative rating data according to the credit rating method. Finally, server 104 can determine the target default probability of the target entity based on the comprehensive rating data, and obtain the credit rating result of the target entity based on the mapping relationship between the target default probability and the credit rating for the target industry exposure, and feed the credit rating result back to terminal 102.

[0050] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0051] In one exemplary embodiment, such as Figure 2 As shown, a credit rating method is provided, which is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 202 to 208. Wherein:

[0052] Step 202: Determine at least one quantitative indicator and at least one qualitative indicator related to the target object and the breach of contract.

[0053] The target entity refers to the subject whose default risk needs to be assessed and whose credit rating needs to be evaluated. Target entities can be various types of entities such as enterprises, individuals, and financial institutions. Default behavior refers to the target entity's failure to fulfill its obligations as agreed, such as an enterprise failing to repay loan principal and interest on time, or an individual failing to repay credit card debt on time. Quantitative indicator data refers to indicator data that can be quantified with specific numerical values. Quantitative indicator data can objectively reflect the characteristics of the target entity related to default behavior. Examples include financial indicators such as a company's debt-to-equity ratio, current ratio, and net profit margin; and an individual's income level and debt amount. Quantitative indicator data allows for intuitive comparison and analysis of the likelihood of default by the target entity through numerical values. Qualitative indicator data refers to indicator data that cannot be precisely measured with specific numerical values ​​and mainly reflects the characteristics of the target entity related to default behavior through subjective judgment and descriptive information. Examples include a company's management level (excellent, good, average, poor, etc.), market reputation (positive, negative, etc.), and industry position (leading, following, etc.); an individual's credit record (whether there are negative credit records, the completeness of the credit record, etc.), job stability (stable, unstable, etc.), and information about the target entity's industry and region.

[0054] For example, in response to a credit rating request for a target object initiated by a user through a terminal, the server can use data acquisition technology, interface calls, database retrieval, manual input verification and other methods to obtain multi-dimensional data such as the target object's financial status, equity structure, industry attributes, regional characteristics, market transaction records, credit information, public opinion dynamics, and support from relevant institutions, and identify at least one quantitative indicator data and at least one qualitative indicator data associated with default behavior from the obtained multi-dimensional data.

[0055] In an optional embodiment, the server can employ data acquisition technology to obtain various raw data, such as the target object, its industry sector, its location, and market transaction records. This data includes information on the target object's financial status, equity structure, industry attributes, regional characteristics, market transaction records, credit information, public opinion dynamics, and support from relevant institutions. By converting financial statement items and indicators across different years, the server can identify and correct any manipulation of the target object's financial statements, thereby obtaining verification results for the authenticity of the financial data and standardized financial indicators that have been corrected. Furthermore, the server can integrate various basic information about the target object, such as credit and credit history data, as well as macroeconomic indicators, to extract the target object's long-term and short-term default characteristics, obtaining default indicators related to the target object's default behavior. These default indicators may include at least one quantitative indicator and at least one qualitative indicator.

[0056] Step 204: Determine the quantitative rating data for at least one quantitative indicator and the qualitative rating data for at least one qualitative indicator.

[0057] Quantitative rating data refers to data obtained after rating quantitative indicator data. Quantitative rating data can be divided into different levels based on certain standards or models to intuitively represent the performance level of a target entity in that quantitative indicator. For example, the debt-to-asset ratio of a target entity can be divided into high, medium, and low levels according to a certain range, with each level corresponding to a quantitative rating data point for subsequent credit evaluation. Another example is determining the target entity's score in asset-related indicators based on its asset-liability ratio to obtain its quantitative rating data. Qualitative rating data refers to data obtained after rating qualitative indicator data. Qualitative rating data can also be obtained according to certain standards or models, transforming the descriptive information of qualitative indicators into a rating format. For example, the management level of a target entity can be classified as excellent, good, average, and poor, each corresponding to different qualitative rating data points, so that it can be comprehensively considered with other indicator data during the credit evaluation process.

[0058] For example, the server can use methods such as grouping and sorting, regression smoothing, monotonicity verification, and standardization to determine the quantitative rating data of at least one quantitative indicator data, and use methods such as graded assignment, expert scoring calibration, and risk level mapping to determine the qualitative rating data of at least one qualitative indicator data.

[0059] Step 206: Based on the target industry exposure of the target object's industry sector, determine the credit evaluation method for the target object, and in accordance with the credit rating method, determine the comprehensive rating data of the target object based on quantitative rating data and qualitative rating data.

[0060] In this context, "target industry exposure" refers to the specific scope or characteristics of the target entity's industry in credit risk assessment. Different industries have different operating characteristics, risk factors, and development patterns. Target industry exposure reflects the degree and manner in which the target entity's industry affects credit risk. For example, high-risk industries (such as real estate and financial derivatives trading) and low-risk industries (such as food processing and public utilities) require different industry factors and weights for credit assessment. "Credit rating method" refers to the specific methods used to conduct credit assessment of the target entity. Depending on the target entity's industry exposure, different credit rating methods can be used to determine its credit rating. In practice, credit rating methods may include, but are not limited to, the evaluation model and indicators used to determine the target entity's credit rating. "Comprehensive rating data" refers to the overall rating result of the target entity obtained by comprehensively analyzing and processing quantitative and qualitative rating data according to the credit rating method, thus comprehensively and accurately reflecting the target entity's credit status. In practice, comprehensive rating data can be presented in the form of grades, such as AAA, AA, A, etc. The higher the grade, the better the credit status of the target and the lower the probability of default. Comprehensive rating data can also be presented in the form of scores, such as 100 points, 80 points, 75 points, etc.

[0061] For example, the server can determine the target industry exposure of the target entity. For instance, the server can combine data such as the target entity's corporate nature, controlling shareholder background, main business type, and core business proportion, and perform cross-validation using preset exposure judgment rules (such as criteria for local government financing vehicles, mapping of financial institution lists, and general industry classification rules) to obtain the target entity's target industry exposure. The server can then determine the credit rating method for the target entity based on its target industry exposure. For example, for exposures with sufficient sample size and complete bad samples, the server can choose a logistic regression model; for exposures with sufficient sample size but missing bad samples, a benchmarking model can be used; for exposures with scarce sample size, an expert experience model can be used, and the indicator weighting rules can be determined based on the risk characteristics of that exposure. Subsequently, the server can determine the target entity's comprehensive rating data based on quantitative and qualitative rating data according to the credit rating method. For example, the server can use the analytic hierarchy process (AHP) combined with expert judgment to determine the weighting of quantitative and qualitative indicators, and then sum the standardized quantitative and qualitative rating data according to their weights to obtain an initial comprehensive score. This score is then mapped to comprehensive rating data through model calibration.

[0062] Step 208: Determine the target default probability of the target entity based on the comprehensive rating data, and obtain the credit rating result of the target entity based on the mapping relationship between the target default probability and the credit rating for the target industry exposure.

[0063] The target default probability (RPP) refers to the likelihood of a target entity defaulting within a future period, as determined by comprehensive rating data. A higher comprehensive rating generally corresponds to a lower RPP, and vice versa. The mapping between RPP and credit rating refers to the correspondence between a pre-defined target RPP and a credit rating for the target industry exposure. This mapping transforms the target RPP into a specific credit rating, providing a more intuitive representation of the target entity's creditworthiness. For example, a target RPP between 0 and 0.1 might correspond to an AAA credit rating, 0.1 to 0.3 to an AA credit rating, and so on. In practice, the mapping between RPP and credit rating can be derived from extensive historical data, industry experience, and statistical analysis. The credit rating result refers to the final credit rating of the target entity for its specific industry exposure, obtained based on the mapping between the target RPP and credit rating.

[0064] For example, the server can determine the target default probability of a target entity based on comprehensive rating data. For instance, the server can use methods such as the K-factor method (suitable for exposures with sufficient default data accumulation) or the parameter traversal method (suitable for low-default portfolio exposures) to determine the target default probability. The server can determine the mapping relationship between the target default probability and the credit rating for the target industry exposure. Then, based on the mapping relationship between the target default probability and the credit rating, the server can obtain the target entity's credit rating result. For example, the server can match the corresponding credit rating range according to the target default probability and the corresponding mapping relationship to obtain the final credit rating result for the target entity.

[0065] In an optional embodiment, when determining the mapping relationship between the target probability of default and the credit rating for the target industry exposure, the server can adjust the default rate range of the external rating master scale. The master scale refers to the standardized mapping relationship between the credit rating of the bond issuer and the probability of default (PD). Then, the server can calibrate and optimize the adjusted external master scale default rate range by combining historical default data and risk characteristics of the target industry exposure, merging them into a dedicated master scale to define the credit ratings (such as AAA, AA+, AA, BBB, etc.) corresponding to different default probability ranges. After obtaining the master scale, the server can substitute the obtained target probability of default into the dedicated master scale to match the corresponding credit rating range, obtaining the final credit rating result.

[0066] The aforementioned credit rating method involves identifying at least one quantitative indicator and at least one qualitative indicator related to the target entity's default behavior; determining the quantitative rating data for each of the at least one quantitative indicator and the qualitative rating data for each of the at least one qualitative indicator; determining the credit evaluation method for the target entity based on its industry exposure; and determining the target entity's comprehensive rating data based on the quantitative and qualitative rating data according to the credit rating method; determining the target entity's target default probability based on the comprehensive rating data; and obtaining the target entity's credit rating result based on the mapping relationship between the target default probability and the credit rating for the target industry exposure. By identifying both quantitative and qualitative indicator data, and separately determining the quantitative rating data for the quantitative indicator and the qualitative rating data for the qualitative indicator, and processing the quantitative and qualitative rating data in conjunction with the target entity's target industry exposure to obtain the target entity's comprehensive rating data, and further determining the target entity's target default probability based on the comprehensive rating data, the target entity's credit rating result can be obtained. This approach comprehensively considers the default-related factors of the target entity and conducts ratings under the corresponding target industry exposure, enhancing the relevance of the rating results and leading to accurate and reliable credit rating results.

[0067] In one embodiment, determining the quantitative rating data for at least one quantitative indicator data and the quantitative rating data for at least one qualitative indicator data respectively includes:

[0068] For each quantitative indicator data, the target object group to which the target object belongs among multiple related object combinations is determined based on the targeted quantitative indicator data. Then, based on the targeted quantitative indicator data, the target object combination, and each related object combination, the quantitative rating data of the targeted quantitative indicator data is determined. The related object combination is obtained by dividing multiple candidate objects that belong to the same business type as the target object into data.

[0069] For each qualitative indicator data, determine the target qualitative interval to which the target qualitative indicator data belongs among multiple associated qualitative intervals, and obtain the qualitative rating data of the target qualitative indicator data based on the mapping relationship between the target qualitative interval and the qualitative data.

[0070] In this context, a "related object portfolio" refers to a set of different objects obtained by segmenting multiple candidate objects belonging to the same business type as the target object. Candidate objects can be similar to the target object in terms of business nature, operating model, market environment, etc. By dividing candidate objects into different portfolios, a comparable reference group can be found when evaluating the quantitative indicators of the target object. For example, when evaluating the debt-to-asset ratio of a bond issuer, all bond issuers of the same type can be used as candidate objects, divided into multiple related object portfolios according to their asset burden. A "target object portfolio" refers to the specific portfolio to which the target object belongs among multiple related object portfolios. By determining the target object portfolio, the target object can be compared with other objects with similar characteristics within the same reference system, thereby more accurately assessing the relative level of its quantitative indicators.

[0071] A related qualitative interval refers to several pre-defined ranges for rating qualitative indicator data. These intervals can be set based on different performance levels or characteristics of the qualitative indicators to categorize the data into corresponding intervals for subsequent rating processing. For example, when assessing the market reputation of a target entity, five related qualitative intervals can be defined: highly positive, positive, neutral, negative, and highly negative. A target qualitative interval refers to the specific interval to which the targeted qualitative indicator data belongs among the multiple related qualitative intervals. By determining the target qualitative interval, the specific performance range of the target entity in terms of the qualitative indicator can be clarified, providing a basis for obtaining qualitative rating data based on the mapping relationship. The mapping relationship between target qualitative intervals and qualitative data refers to the correspondence between the pre-defined target qualitative intervals and the final qualitative rating data. This mapping relationship clarifies the qualitative rating data corresponding to each target qualitative interval, allowing the corresponding qualitative rating data to be directly obtained by determining the target qualitative interval. For example, we can define the qualitative rating data corresponding to a highly positive target qualitative interval as A, positive as B, and so on. Or, we can define the qualitative rating data corresponding to a highly positive target qualitative interval as 100 points, positive as 80 points, and so on.

[0072] For example, for each quantitative indicator data, the server can determine the target object group to which the target object belongs among multiple related object combinations based on the targeted quantitative indicator data. For instance, the server can obtain multiple related object combinations by sorting the quantitative indicator values ​​from high to low or low to high, grouping by equal frequency, or grouping by equal intervals. The server can further match the quantitative indicator data of the target object with each related object combination to obtain the target object group to which the target object belongs among multiple related object combinations. The server can determine the quantitative rating data of the targeted quantitative indicator data based on the targeted quantitative indicator data, the target object combination, and each related object combination; for example, the server can determine the default rate of the target object combination, perform smooth fitting and monotonicity verification through regression, and then obtain the quantitative rating data corresponding to the quantitative indicator data. For each qualitative indicator data point, the server can determine the target qualitative interval to which the target qualitative indicator data belongs among multiple related qualitative intervals. For example, the server can pre-set multi-dimensional grading standards based on the economic meaning and business logic of the qualitative indicator (such as support level and business compliance) and combine them with expert experience verification to obtain multiple related qualitative intervals. The server then matches the qualitative indicator data of the target object with each related qualitative interval to obtain the target qualitative interval to which the target object's qualitative indicator data belongs. Afterward, the server can obtain the qualitative rating data for the target qualitative indicator data based on the mapping relationship between the target qualitative interval and the qualitative data. For example, the server can configure corresponding scores for each related qualitative interval, such as 100 points, 80 points, etc., and use the score matched by the target qualitative interval as the qualitative rating data for that qualitative indicator data.

[0073] In this embodiment, for quantitative indicator data, quantitative rating data is obtained by determining the target object combination to which the target object belongs in the associated object combination. For qualitative indicator data, qualitative rating data is obtained by determining the target qualitative interval to which it belongs in the associated qualitative interval and by using the mapping relationship. This can improve the scientificity and pertinence of qualitative and quantitative evaluation of the target object and help provide reliable basic data for subsequent data processing.

[0074] In one embodiment, determining the target object group to which the target object belongs among multiple related object groups based on the targeted quantitative indicator data includes:

[0075] Obtain the quantitative indicator data of multiple candidate objects, including the target object, that belong to the same business type as the target object; divide each candidate object into multiple related object combinations according to the data distribution characteristics of each quantitative indicator data; match the quantitative indicator data of the target object with the quantitative indicator data range of each related object combination to obtain the target object combination to which the target object belongs.

[0076] Data distribution characteristics refer to the distribution patterns of quantitative indicator data within a candidate object group. These characteristics may include, but are not limited to, central tendency (e.g., mean, median, mode), dispersion (e.g., standard deviation, variance, range), and distribution shape (e.g., normal distribution, skewed distribution). Different data distribution characteristics can be divided using different methods and standards to ensure the rationality and effectiveness of the resulting combinations of related objects. For example, for distribution characteristics that are approximately normal or relatively uniform with no significant extreme values, candidate objects can be divided according to the equal frequency of the corresponding quantitative indicator data from high to low, resulting in multiple combinations of related objects.

[0077] For example, the server can acquire quantitative indicator data for multiple candidate objects, including the target object, that belong to the same business type as the target object. For instance, the server can collect historical and real-time data on quantitative indicators such as asset-liability ratio, cash flow coverage ratio, and net profit growth rate of candidate objects in the same industry (e.g., urban investment companies, manufacturing enterprises) through industry databases, internal credit systems of financial institutions, and synchronization with credit reporting platform interfaces. After obtaining the quantitative indicator data, the server can divide each candidate object into multiple related object combinations according to the data distribution characteristics of each quantitative indicator. For example, the server can use an equal-frequency grouping method for normally distributed quantitative indicators and an equal-interval grouping method for skewed distribution indicators or indicators with extreme values, dividing the candidate objects into multiple related object combinations with clearly defined intervals. Afterwards, the server can match the quantitative indicator data of the target object with the quantitative indicator data range of each associated object combination to obtain the target object combination to which the target object belongs. For example, the server can extract the specific value of the quantitative indicator of the target object (such as a debt-to-asset ratio of 75%) and compare it with the indicator range of each associated object combination (such as the [70%-80%] group). If the value falls within the range of a certain group, then the group is determined to be the target object combination to which the target object belongs.

[0078] In this embodiment, by acquiring quantitative indicator data of multiple candidate objects of the same business type, and dividing related object combinations according to the data distribution characteristics of the quantitative indicator data of each candidate object, and then matching the quantitative indicator data of the target object with the range of quantitative indicator data in each related object combination, the related object combinations can be comprehensively and reasonably divided, thereby accurately determining the target object combination to which the target object belongs. This provides a reliable foundation for subsequent analysis based on combination characteristics and improves the accuracy and efficiency of business processing.

[0079] In one embodiment, such as Figure 3 As shown, based on the targeted quantitative indicator data, the target object combination, and each related object combination, the quantitative rating data for the targeted quantitative indicator data is determined, including:

[0080] Step 302: For each associated object combination, determine the combined quantitative data of the associated object combination based on the quantitative indicator data of each candidate object included in the associated object combination.

[0081] In this context, portfolio quantitative data refers to the quantitative indicator data of each candidate object within a portfolio of related objects. This data reflects the overall level of the portfolio in terms of a specific quantitative indicator, providing a basis for subsequent correlation analysis with the portfolio default rate. In practice, portfolio quantitative data can be a comprehensive set of data determined using methods such as average, median, or weighted average. For example, for a portfolio of related objects consisting of multiple candidate objects, if the target quantitative indicator is the debt-to-equity ratio, then the portfolio quantitative data could be the average of the debt-to-equity ratios of all candidate objects within that portfolio.

[0082] For example, for each associated object combination, the server can determine the combined quantitative data of the associated object combination based on the quantitative indicator data of each candidate object included in the associated object combination; for example, the server can determine the mean, median or quantile (such as the 75th quantile) of the quantitative indicator data of all candidate objects in the associated object combination, and use the statistical value as the combined quantitative data of the associated object combination to characterize the overall indicator level of the group.

[0083] Step 304: Determine the combined default rate of the combined associated objects based on the individual default rates of each candidate object included in the combined associated objects.

[0084] The individual default rate (IFR) refers to the probability of a candidate entity defaulting, reflecting the likelihood of such default. In practice, the individual default rate can be obtained by statistically analyzing the historical default data of candidate entities. The portfolio default rate (BFR) refers to the overall credit risk level of the portfolio of related entities, based on the individual default rates of each candidate entity within the portfolio. In practice, the BFR is the overall probability of default for the portfolio, determined using a method such as weighted averaging. For example, if a portfolio has three candidate entities with individual default rates of 5%, 10%, and 15%, a weighted average of these rates yields the portfolio's BFR.

[0085] For example, the server can determine the combined default rate of the associated object combination based on the default rates of each candidate object included in the associated object combination. For instance, the server can determine the original combined default rate by dividing the number of defaulted samples in the combination by the total number of samples in the combination. If the sample size is small, the final combined default rate can be obtained by smoothing it by adding the industry average default rate.

[0086] In an optional embodiment, the server can compare the default rate levels of missing groups and non-missing groups and place the missing group in an appropriate position. For example, the server can determine the default rate of the missing group composed of candidate objects with missing quantitative indicator data, and compare the difference between the default rate of the missing group and the default rate of each non-missing group. If the default rate of the missing group is between a certain group of associated objects, the server can insert the missing group between the two groups of associated objects to form a new group. If the default rate of the missing group falls into a certain group of associated objects, the server can add the missing group to the corresponding group of associated objects to update the previous grouping.

[0087] Step 306: If the quantitative data and default rate of each portfolio meet the trend change conditions, obtain the quantitative rating data of the targeted quantitative indicator data based on the default rate of each portfolio.

[0088] Trend change conditions can be used to determine whether a specific trend relationship exists between portfolio quantitative data and portfolio default rate. For example, trend change conditions could be that as portfolio quantitative data increases, the portfolio default rate also shows an upward trend; or that when portfolio quantitative data changes within a certain range, the change in portfolio default rate conforms to a pre-defined functional relationship. Trend change conditions can ensure a reasonable correlation between portfolio quantitative data and portfolio default rate. When portfolio quantitative data and portfolio default rate meet trend change conditions, reliable quantitative rating data can be obtained based on the portfolio default rate.

[0089] For example, the server can obtain trend change conditions, such as a monotonic correlation between portfolio quantitative data and portfolio default rates (e.g., positive or negative correlation, consistent with economic meaning and business logic; for example, a higher debt-to-equity ratio corresponds to a higher portfolio default rate; a higher current ratio corresponds to a lower portfolio default rate). Then, the server can use a loss function (e.g., locally weighted regression with scatter smoothing) to fit the portfolio quantitative data and portfolio default rates of all related object portfolios, verifying whether the fitted trend meets the monotonic correlation requirement, determining whether the portfolio quantitative data and portfolio default rates meet the trend change conditions, and if the quantitative data and default rates of each portfolio meet the trend change conditions, the server can obtain quantitative rating data for the targeted quantitative indicator based on the default rates of each portfolio. For example, the server can divide the portfolio default rate of the target object's portfolio by the weighted average of the portfolio default rates of all related object portfolios (the weight being the proportion of candidate objects in each related object portfolio) to obtain standardized quantitative rating data; a higher value indicates a higher default risk corresponding to that indicator.

[0090] In an optional embodiment, the server can also use Loss regression (locally weighted regression scatter smoothing method) to smooth the data and ensure the robustness of the default rates of each group. For example, the server can construct a scatter plot with the combined quantitative data (such as the average debt-to-equity ratio) of each related object combination as the x-axis and the combined default rate as the y-axis. By using the Loss regression algorithm to give higher weights to neighboring samples for local fitting, the server can correct the outlier in the original data where the default rate of the high index value group is lower than that of the low index value group, making the default rate trend more logical.

[0091] In an optional embodiment, for some non-monotonic or abrupt situations, the server can determine whether the phenomenon conforms to economic meaning. For example, the server can analyze whether there is a reasonable reason for the non-monotonic situation where the higher the current ratio (short-term solvency indicator) is, the higher the portfolio default rate (e.g., some companies hold excessive cash leading to a decline in profits), and verify the rationality of the phenomenon in combination with industry characteristics and business logic. If it does not conform, the server can force the variable to conform to the trend of economic meaning. For example, the server can use the mean of the default rate of adjacent groups to replace the abnormally high default rate of 2.8% of the current ratio with the mean of the two adjacent groups, ensuring that the current ratio and the default rate have a negative correlation trend. If it conforms, the server can force monotonicity if the fitted data can better support the business logic. For example, the server can constrain the sign of the derivative of the fitted function to correct the slight jumps of 3.2% in the default rate of the 70%-80% current ratio group and 3.0% in the default rate of the 80%+ current ratio group to 3.2% and 3.5%, respectively, maintaining the monotonically increasing trend of the default rate as the indicator value increases.

[0092] In this embodiment, by determining the combined quantitative data and combined default rate of the associated object group, the overall situation and default risk level of the candidate objects in each associated object group on specific quantitative indicator data can be comprehensively reflected. When the combined quantitative data and combined default rate meet the trend change conditions, quantitative rating data is obtained based on the combined default rate. This can be combined with the quantitative indicator data itself and the default risk factors of the candidate objects, making the quantitative rating more scientific, comprehensive and accurate, and reducing the risk and uncertainty of the evaluation process.

[0093] In one embodiment, determining the target default probability of a target entity based on comprehensive rating data includes:

[0094] Determine the mapping relationship between comprehensive rating data and default probability, and multiple mapping coefficients for the mapping relationship; for each mapping coefficient, determine the candidate default probability of the target object based on the comprehensive rating data, the mapping relationship, and the corresponding mapping coefficient; determine the target default probability of the target object from the candidate default probabilities based on the industry default probability of the industry sector to which the target object belongs.

[0095] The probability of default refers to the likelihood that a target entity will default within a specific future timeframe, and it can be expressed as a percentage. The mapping relationship between comprehensive rating data and the probability of default describes the corresponding default probability for different levels of comprehensive rating data. This mapping relationship can be determined through statistical analysis of historical data and the establishment of mathematical models (such as logistic regression, decision tree, and neural network models). For example, analyzing historical data reveals that when the comprehensive rating score is between 80 and 90, the probability of default is typically between 2% and 5%, thus establishing a mapping relationship between the comprehensive rating data and the probability of default. Alternatively, a mathematical model can be used to determine this mapping relationship. Mapping coefficients are parameters used to adjust the mapping relationship, refining and correcting it. Different mapping coefficients can adjust the mapping relationship from different perspectives.

[0096] Candidate default probability refers to the probability of a target object potentially defaulting, determined using a specific method based on comprehensive rating data, mapping relationships, and the specific mapping coefficient, for each mapping coefficient. Since multiple sets of mapping coefficients exist, multiple candidate default probabilities can be obtained. For example, assuming three different sets of mapping coefficients, three different candidate default probabilities will be obtained based on these three sets. Industry default probability refers to the average probability of all objects within the target object's industry sector defaulting within a specific time period. Industry default probability can be obtained through statistical analysis of historical default data within the industry.

[0097] For example, the server can determine the mapping relationship between comprehensive rating data and default probability, as well as multiple mapping coefficients for the mapping relationship. For instance, the server can use a linear regression analysis model to construct the mapping relationship (the model expression is ln(PD / (1-PD))=α+β×Score, where PD is the default probability, Score is the comprehensive rating data, and α and β are the mapping coefficients). Based on this, the server can preset the range of the mapping coefficient α to [-5,-2] and the range of the mapping coefficient β to [1.5,3.0], generating multiple combinations of α and β as candidate mapping coefficients. For each candidate mapping coefficient, the server can determine the candidate default probability of the target object based on the comprehensive rating data, the mapping relationship, and the corresponding mapping coefficient. For example, the server can substitute the comprehensive rating data of the target object (e.g., 95 points) into the above linear regression analysis model, and combine it with a certain set of candidate mapping coefficients (e.g., α=-3.5, β=2.0) to determine the candidate default probability PD≈12.5%. Then, the server can determine the target default probability of the target object from the candidate default probabilities based on the industry default probability of the target object's industry sector. For example, the server can obtain the average default rate (e.g., 2.5%-3.5%) of the target industry exposure (such as local government financing vehicles) over the past 3 years as the industry default probability, determine the degree of deviation between each candidate default probability and the industry default probability, screen the candidate default probability with the smallest deviation and that meets the industry risk characteristics, or perform weighted calibration of the candidate default probabilities through the industry default probability, and finally determine the target default probability.

[0098] In this embodiment, by determining the mapping relationship between comprehensive rating data and default probability and multiple mapping coefficients, the candidate default probability of the target object is determined for each mapping coefficient. Combined with the industry default probability of the target object's industry, the target default probability is determined from the candidate default probabilities. This makes the determination of the target default probability more realistic and effectively improves the accuracy and reliability of default risk assessment.

[0099] In one embodiment, the credit rating methodology further includes:

[0100] Identify at least one of the following: credit change information, external correlation information, and historical rating deviation information of the target entity; revise the credit rating result based on at least one of the following: credit change information, external correlation information, and historical rating deviation information, to obtain the revised credit rating result.

[0101] Credit change information reflects data and circumstances that indicate changes in the creditworthiness of the target entity over time. For example, it includes information on situations that occur infrequently but, if they do occur, would significantly deteriorate the credit quality of the counterparty and the issuer. External association information refers to information provided by other entities or external factors related to the target entity. For instance, it includes information on support or assistance that related institutions or the parent company might provide in the event of an emergency. Therefore, related institutions can make corresponding adjustments based on the rating results. When rating the parent company, group consolidation can be used, and the existence of written joint guarantees or other implicit guarantees can be considered to make corresponding adjustments. Historical rating deviation information refers to the discrepancy between the actual rating results when the target entity was previously rated and its subsequent actual credit performance.

[0102] For example, the server can determine at least one of the following: credit change information, external correlation information, and historical rating deviation information of the target entity. For instance, the server can use methods such as monitoring public opinion to capture sudden negative events affecting the target entity (e.g., illegal debt issuance, project failure), synchronizing the latest financial statements and credit overdue records via interfaces, and retrieving historical rating results from its own rating system for comparison with actual defaults. These methods allow the server to determine the target entity's credit change information (e.g., a sharp drop in recent operating cash flow, increased debt concentration), external correlation information (e.g., adjustments to local government financial subsidy policies, changes in the parent company's joint guarantee liability), and historical rating deviation information (e.g., past rating results underestimated actual risk by several levels). The server can then correct the credit rating result based on at least one of these three information. For example, if significant violations are detected in the target entity's public opinion (credit change information), the server can lower the initial rating by one level; if local government issues special support policies (external correlation information), the server can raise the rating by one level. Simultaneously, the server can adjust the current PD value using a coefficient based on the historical rating deviation rate, ultimately integrating all corrective factors to obtain the corrected credit rating result.

[0103] In this embodiment, by acquiring at least one of the following: credit change information of the target object, external correlation information, and historical rating deviation information, the factors affecting the credit rating can be comprehensively determined from multiple key dimensions. Based on this information, the credit rating result can be corrected, which can effectively make up for the shortcomings of the initial rating, so that the corrected credit rating result is more in line with the true credit status of the target object and enhance the accuracy and reliability of the credit rating.

[0104] In one application example, such as Figure 4 As shown, a credit rating method is provided, which specifically includes the following steps:

[0105] Step 1: Data extraction.

[0106] Data acquisition technology acquires various data such as enterprise, industry, region, and market transactions; it can identify financial statement manipulation by converting financial statement items and indicators between different years; and it integrates basic information of various entities such as credit and credit reporting within the industry, as well as macroeconomic indicators, to extract long-term and short-term default characteristics (i.e., quantitative and qualitative indicator data).

[0107] Step 2: Data processing.

[0108] For quantitative indicator data, variables (i.e., quantitative indicator data of the target object and other related candidate objects) can be sorted in a certain order (from high to low or from low to high), and the sample can be divided into several groups in turn to obtain the grouping status of each variable, i.e., the combination of related objects; compare the default rate levels of the missing group and the non-missing group, and place the missing group in an appropriate position, i.e., determine the target object combination; use Loss regression (local weighted regression scatter smoothing method) to smooth the data and ensure the robustness of the default rate of each group; for some non-monotonic or jump cases, determine whether the phenomenon conforms to economic meaning. If it does not, the variable is forced to conform to the trend of economic meaning; if it does, the fitted data can better support the business logic, and then it is forced to be monotonic; divide the default rate corresponding to each group by the weighted average after monotonicity processing, and perform standardization processing to obtain quantitative rating data.

[0109] For qualitative indicator data, scores can be set for each group to replace the indicator grading for each group (i.e., the associated qualitative interval) to obtain qualitative rating data.

[0110] Step 3: Opening division.

[0111] The exposure of urban investment companies is determined based on the target company's nature, major shareholders, industry attributes, financial data, etc.; the exposure of financial institutions is mapped according to the exposure list of each financial institution; the exposure of general industries is mapped according to general industry mapping rules; if it is a cross-industry group or an institution without a clear scorecard, a general corporate scorecard is used to cover it. Since its industry attributes are not clear, it is handled by rating superposition or adjustment.

[0112] Step 4: Initial model construction.

[0113] A modeling dataset is constructed based on company financial data, annual reports, bond issuance prospectuses, and rating reports. Single-factor analysis is used to screen qualitative and quantitative indicators capable of differentiating company credit ratings within each exposure, and these indicators are adjusted and confirmed using expert experience. Based on accumulated industry experience and expert judgments on the importance of indicators, and guided by expert opinion, the weights of quantitative and qualitative indicators are determined through statistical methods. The model is determined based on sample quality: for exposures with sufficient sample size and a relatively large number of bad samples, a logistic regression model is used to classify each institution as good or bad annually; for exposures with sufficient sample size but missing bad samples, the risk level of institutions is ranked according to expert experience, and a benchmarking model is used; for exposures with insufficient samples, the model is determined based on expert experience.

[0114] Step 5: Model calibration.

[0115] To determine the relationship between model scores and probability of default (PD), that is, to find suitable parameters to determine the PD value for each sample. For banking financial institutions with a certain amount of default data accumulation, the K-factor method is used for model calibration, which is divided into three steps: determining the initial default probability, determining the K factor (i.e., mapping coefficients), and calculating the actual default overview. For exposures of low-default asset portfolios, the parameter traversal method is used to select the parameter combination that best matches the calibration result requirements, that is, to obtain the total score (score) of all samples for each model exposure. The calibration parameters α and β are traversed, and for each combination of α and β, the default probability PD of each target object is determined according to the formula ln(PD / (1-PD))=α+β×Score (where PD is the default probability, Score is the comprehensive rating data, and α and β are mapping coefficients), and then the rating level of the customer is obtained through the main scale mapping. In addition, the default probability obtained from each parameter combination can be filtered using the industry default probability to obtain the final target default probability.

[0116] Step 6: Establish the master scale.

[0117] The primary benchmark maps the issuer's credit rating to its default profile. The default rates of external rating primary benchmarks are adjusted; these external rating primary benchmarks are then merged to form the primary benchmark for this credit rating methodology. After obtaining the target default probability, this target default probability can be compared with the primary benchmark to arrive at the corresponding credit rating.

[0118] Furthermore, after obtaining the credit rating of the target entity, at least one of the following may also be included:

[0119] Step 7: External adjustment of the model.

[0120] For situations that occur infrequently but, once they do occur, would significantly deteriorate the credit quality of the counterparty and the issuer, out-of-model adjustment rules are set up to make corresponding adjustments. If an institution encounters an emergency, the relevant institution or its parent company is likely to provide support or rescue; therefore, for the relevant institution, corresponding adjustments are made based on the above rating results. When rating the parent company, group consolidation is adopted, and the existence of written joint and several guarantees or other implicit guarantees is considered to make corresponding adjustments.

[0121] Step 8: Model overturning.

[0122] Review the model's history of revisions; determine if the reasons for revisions are not risk factors covered by the model, and based on the principle of prudence, identify strong evidence for revision; make adjustments to the existing rating within a certain range.

[0123] Based on the above process, this example combines multi-dimensional data such as financial data after window dressing identification, basic information of the entity, industry / regional data, market transaction data, industry credit / credit data, and macroeconomic data. It combines multiple methods to apply to the credit rating of companies with different exposures. By calibrating the model and establishing a master benchmark, it merges the credit rating results of entities with different exposures and considers the support factors of relevant institutions and parent companies for out-of-model adjustments, which can effectively improve the interpretability and accuracy of the rating results.

[0124] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0125] Based on the same inventive concept, this application also provides a credit rating apparatus for implementing the credit rating method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more credit rating apparatus embodiments provided below can be found in the limitations of the credit rating method described above, and will not be repeated here.

[0126] In one exemplary embodiment, such as Figure 5 As shown, a credit rating device is provided, including: a data determination module 502, an indicator rating module 504, a comprehensive rating module 506, and a credit rating module 508, wherein:

[0127] The data determination module 502 is used to determine at least one quantitative indicator data and at least one qualitative indicator data related to the target object and the breach of contract.

[0128] The indicator rating module 504 is used to determine the quantitative rating data of at least one quantitative indicator data and the qualitative rating data of at least one qualitative indicator data respectively.

[0129] The comprehensive rating module 506 is used to determine the credit evaluation method for the target object based on the target industry exposure of the target object's industry sector, and to determine the comprehensive rating data of the target object based on quantitative rating data and qualitative rating data according to the credit rating method.

[0130] Credit rating module 508 is used to determine the target default probability of the target entity based on comprehensive rating data, and to obtain the credit rating result of the target entity based on the mapping relationship between the target default probability and the credit rating for the target industry exposure.

[0131] In an optional embodiment, the indicator rating module 504 is further configured to, for each quantitative indicator data, determine the target object combination to which the target object belongs among multiple related object combinations based on the quantitative indicator data, and determine the quantitative rating data of the quantitative indicator data based on the quantitative indicator data, the target object combination, and each related object combination; the related object combination is obtained by dividing multiple candidate objects that belong to the same business type as the target object; for each qualitative indicator data, determine the target qualitative interval to which the qualitative indicator data belongs among multiple related qualitative intervals, and obtain the qualitative rating data of the qualitative indicator data based on the mapping relationship between the target qualitative interval and the qualitative data.

[0132] In an optional embodiment, the indicator rating module 504 is further configured to obtain the quantitative indicator data of multiple candidate objects, including the target object and belonging to the same business type as the target object; divide each candidate object into multiple associated object combinations according to the data distribution characteristics of each quantitative indicator data; and match the quantitative indicator data of the target object with the quantitative indicator data range of each associated object combination to obtain the target object combination to which the target object belongs.

[0133] In an optional embodiment, the indicator rating module 504 is further configured to, for each associated object combination, determine the combined quantitative data of the associated object combination based on the targeted quantitative indicator data of each candidate object included in the associated object combination; determine the combined default rate of the associated object combination based on the object default rate of each candidate object included in the associated object combination; and, if the combined quantitative data and the combined default rate meet the trend change conditions, obtain the quantitative rating data of the targeted quantitative indicator data based on the combined default rate.

[0134] In an optional embodiment, the credit rating module 508 is further configured to determine the mapping relationship between the comprehensive rating data and the default probability, and multiple mapping coefficients for the mapping relationship; for each mapping coefficient, based on the comprehensive rating data, the mapping relationship and the mapping coefficient, determine the candidate default probability of the target object; and determine the target default probability of the target object from the candidate default probabilities according to the industry default probability of the industry sector to which the target object belongs.

[0135] In an optional embodiment, the credit rating device further includes a rating correction module for determining at least one of the target object's credit change information, external correlation information, and historical rating deviation information; and correcting the credit rating result based on at least one of the credit change information, external correlation information, and historical rating deviation information to obtain a corrected credit rating result.

[0136] Each module in the aforementioned credit rating device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0137] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores qualitative and quantitative indicator data, credit rating methods, and various mapping relationships. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a credit rating method.

[0138] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0139] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the credit rating methods of the above embodiments.

[0140] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the credit rating methods of the above embodiments.

[0141] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the credit rating methods of the above embodiments.

[0142] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0143] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0144] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0145] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A credit rating method, characterized in that, The method includes: Identify at least one quantitative indicator and at least one qualitative indicator that link the target entity to the breach of contract. Determine the quantitative rating data for each of the at least one quantitative indicator data and the qualitative rating data for each of the at least one qualitative indicator data; Based on the target industry exposure of the target object to which the target object belongs, determine the credit evaluation method for the target object, and in accordance with the credit rating method, determine the comprehensive rating data of the target object based on the quantitative rating data and the qualitative rating data; Based on the comprehensive rating data, the target default probability of the target object is determined, and based on the mapping relationship between the target default probability and the credit rating for the target industry exposure, the credit rating result of the target object is obtained.

2. The method according to claim 1, characterized in that, The step of determining the quantitative rating data for each of the at least one quantitative indicator data and the quantitative rating data for each of the at least one qualitative indicator data includes: For each quantitative indicator data, the target object group to which the target object belongs in multiple associated object combinations is determined based on the targeted quantitative indicator data, and the quantitative rating data of the targeted quantitative indicator data is determined based on the targeted quantitative indicator data, the target object group, and each of the associated object groups; the associated object group is obtained by dividing multiple candidate objects that belong to the same business type as the target object into data; For each qualitative indicator data, a target qualitative interval is determined among multiple associated qualitative intervals to which the target qualitative indicator data belongs, and qualitative rating data for the target qualitative indicator data is obtained based on the mapping relationship between the target qualitative interval and the qualitative data.

3. The method according to claim 2, characterized in that, The step of determining the target object combination to which the target object belongs among multiple related object combinations based on the targeted quantitative indicator data includes: Obtain quantitative indicator data for each of multiple candidate objects, including the target object and belonging to the same business type as the target object; Based on the data distribution characteristics of each of the quantitative indicator data, each of the candidate objects is divided into multiple related object combinations; The quantitative indicator data of the target object are matched with the quantitative indicator data range of each of the associated object combinations to obtain the target object combination to which the target object belongs.

4. The method according to claim 2, characterized in that, The determination of quantitative rating data for the targeted quantitative indicator data based on the targeted quantitative indicator data, the target object combination, and each of the associated object combinations includes: For each group of associated objects, the combined quantitative data of the group of associated objects is determined based on the quantitative index data of each candidate object included in the group of associated objects. The combined default rate of the targeted associated object combination is determined based on the individual default rates of each candidate object included in the targeted associated object combination. When the combined quantitative data and the combined default rate meet the trend change conditions, the quantitative rating data of the targeted quantitative indicator data is obtained based on the combined default rate.

5. The method according to claim 1, characterized in that, Determining the target default probability of the target object based on the comprehensive rating data includes: Determine the mapping relationship between the comprehensive rating data and the probability of default, as well as multiple mapping coefficients for the mapping relationship; For each mapping coefficient, based on the comprehensive rating data, the mapping relationship, and the mapping coefficient, the candidate default probability of the target object is determined; Based on the industry default probability of the target object's industry sector, the target default probability is determined from the candidate default probabilities.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Determine at least one of the following: credit change information, external association information, and historical rating deviation information of the target object; The credit rating result is corrected based on at least one of the credit change information, the external correlation information, and the historical rating deviation information to obtain the corrected credit rating result.

7. A credit rating device, characterized in that, The device includes: The data determination module is used to determine at least one quantitative indicator data and at least one qualitative indicator data that are associated with the target object and the breach of contract. The indicator rating module is used to determine the quantitative rating data of each of the at least one quantitative indicator data and the qualitative rating data of each of the at least one qualitative indicator data. The comprehensive rating module is used to determine the credit rating method for the target object based on the target industry exposure of the target object's industry sector, and to determine the comprehensive rating data of the target object based on the quantitative rating data and the qualitative rating data according to the credit rating method. The credit rating module is used to determine the target default probability of the target object based on the comprehensive rating data, and to obtain the credit rating result of the target object based on the mapping relationship between the target default probability and the credit rating for the target industry exposure.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images