Personal business loan credit assessment method based on bayesian learning and related products

Abstract

The application provides a personal business loan credit evaluation method based on Bayesian learning and related products, wherein the implementation of the method comprises: receiving a loan application from a target user for requesting a guarantee loan for a target business subject; obtaining business plan information of the target business subject, credit evaluation information of the target user, and historical business condition information of other business subjects under the target user; analyzing the business correlation between the business plan information of the target business subject and the historical business condition information of the other business subjects to determine a first business risk value; inputting the credit evaluation information of the target user and the business plan information of the target business subject into a Bayesian learning model to obtain a second business risk value; and determining an evaluation result of the loan application according to the first business risk value and the second business risk value. The method of the application embodiment can accurately evaluate the credit of personal business loans, thereby avoiding loan risks.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a personal business loan credit evaluation method based on Bayesian learning and related products. BACKGROUND

[0002] With the rapid development of social economy and the continuous improvement of productivity, market business activities are increasingly active. In order to promote the further development of economy to meet the demand of market business entities for expanding productivity for funds, the loan guarantee behavior for the condition of repayment and interest payment also presents an explosive growth in quantity.

[0003] Among the large number of loan guarantee behaviors, there are several different loan categories, and the personal business loan is one of them. When a financial institution issues a personal business loan, it needs to understand the borrower's own situation and the operation of the borrower's business in detail. Due to the large amount of information used for credit evaluation, it is difficult for the financial institution to conduct comprehensive and accurate credit evaluation on the personal business loan, so the risk control of the personal business loan is more difficult, which leads to a certain loan risk when the financial institution issues the personal business loan. SUMMARY

[0004] The embodiments of the present application provide a personal business loan credit evaluation method based on Bayesian learning and related products. By implementing the embodiments of the present application, the personal business loan is accurately evaluated, and the loan risk is avoided.

[0005] In a first aspect, the embodiments of the present application provide a personal business loan credit evaluation method based on Bayesian learning. The method comprises:

[0006] Receiving a loan application from a target user, the loan application being used to request a guaranteed loan for a target business entity;

[0007] Obtaining business plan information of the target business entity, credit evaluation information of the target user, and historical business condition information of other business entities under the target user;

[0008] Analyzing the business correlation between the business plan information of the target business entity and the historical business condition information of the other business entities, and determining a first business risk value;

[0009] Inputting the credit evaluation information of the target user and the business plan information of the target business entity into the Bayesian learning model as input items, and obtaining a second business risk value;

[0010] According to the first business risk value and the second business risk value, an evaluation result of the loan application is determined, and the evaluation result of the loan application includes passing the loan application or not passing the loan application.

[0011] In one possible example, the user evaluation condition includes a first condition, and the first condition represents that the number of intellectual property under the historical user name is greater than a preset number, and the business entity evaluation condition includes a second condition, and the second condition represents that the R&D expense proportion of the business entity is greater than a preset proportion, and at least one combined condition event is obtained according to at least one user evaluation condition and at least one business entity evaluation condition, including:

[0012] The combined condition event is obtained according to occurrence of the first condition and occurrence of the second condition, and / or the combined condition event is obtained according to non-occurrence of the first condition and non-occurrence of the second condition.

[0013] In a second aspect, an embodiment of the present application provides a personal business loan credit assessment device based on Bayesian learning, and the device includes:

[0014] A receiving unit is configured to receive a loan application from a target user, and the loan application is used to request a guaranteed loan for a target business entity.

[0015] An obtaining unit is configured to obtain business plan information of the target business entity, credit assessment information of the target user, and historical business status information of other business entities under the target user name.

[0016] An analyzing unit is configured to analyze business correlation between the business plan information of the target business entity and the historical business status information of the other business entities, and determine a first business risk value.

[0017] A model unit is configured to input the credit assessment information of the target user and the business plan information of the target business entity into a Bayesian learning model as input items, and obtain a second business risk value.

[0018] An evaluation unit is configured to determine an evaluation result of the loan application according to the first business risk value and the second business risk value, and the evaluation result of the loan application includes passing the loan application or not passing the loan application.

[0019] In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, and computer-executable instructions stored on the memory and executable on the processor, when the computer-executable instructions are executed, causing the electronic device to perform part or all of the steps described in any method of the first aspect of the present application.

[0020] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores computer instructions, and when the computer instructions run on a communication device, the communication device executes part or all of the steps described in any method of the first aspect of the embodiments of the present application.

[0021] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a computer program, and the computer program is operable to cause a computer to execute part or all of the steps described in any method of the first aspect of the embodiments of the present application. The computer program product can be a software installation package.

[0022] It can be seen that, in the embodiments of the present application, the first operating risk value is determined by analyzing the operating relevance between the business plan information of the target business subject and the historical operating condition information of other business subjects, the credit assessment information of the target user and the business plan information of the target business subject are input into the Bayesian learning model as input items, the second operating risk value is obtained, and the evaluation result of the loan application is determined according to the first operating risk value and the second operating risk value. When facing the information of users and business subjects in many aspects and in large quantities, the financial institution can comprehensively and accurately perform credit assessment on the personal operating loan by using the method of the embodiments of the present application, and thus the loan risk can be avoided and the sustainable development of the loan business can be ensured. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0024] Figure 1 is an example schematic diagram of a credit assessment system;

[0025] Figure 2 is a flowchart of a personal operating loan credit assessment method based on Bayesian learning provided by an embodiment of the present application;

[0026] Figure 3 is a structural schematic diagram of a Bayesian learning model provided by an embodiment of the present application;

[0027] Figure 4 is an architectural schematic diagram of a personal operating loan credit assessment system based on Bayesian learning provided by an embodiment of the present application;

[0028] Figure 5is an example schematic diagram of a personal business loan credit assessment method based on Bayesian learning provided by an embodiment of the present application.

[0029] Figure 6 is a structural schematic diagram of a personal business loan credit assessment device based on Bayesian learning provided by an embodiment of the present application.

[0030] Figure 7 is a server structure schematic diagram of a hardware running environment of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0031] In order for those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0032] The terms "first", "second", and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish different objects, not to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps is not limited to the listed steps, but can optionally include steps not listed, or can optionally include other steps inherent to the process, method, product or device.

[0033] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor does it necessarily refer to a separate or alternative embodiment in isolation or in combination with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0034] The application scenarios related to the embodiments of the present application will be introduced below with reference to the drawings.

[0035] Figure 1 is an example schematic diagram of a credit assessment system. As shown in Figure 1 , the system includes a target user terminal, a financial institution terminal and a risk assessment model.

[0036] Among them, the target user terminal refers to a role with loan guarantee needs, thereby submitting a loan application to the financial institution terminal to request to obtain a guaranteed loan, and the loan application includes credit assessment information of the target user;

[0037] The credit assessment information can include credit assessment information of the target user, and when the guarantee loan is a personal business loan, the credit assessment information can further include business plan information or other information of the business subject to be invested by the target user.

[0038] The financial institution terminal receives the loan application from the target user terminal, inputs the credit assessment information in the loan application into the risk assessment model to obtain the evaluation result of the target user, and determines whether to issue the guarantee loan to the target user according to the evaluation result.

[0039] The risk assessment model is used to evaluate the credit assessment information input therein, determine the loan default risk corresponding to the credit assessment information, and output the evaluation result.

[0040] It can be seen that in the process of credit assessment of the above system, since the risk assessment model only uses the traditional evaluation mechanism to predict the loan default risk corresponding to the credit assessment information, when the loan type applied by the target user is a personal business loan with a large amount of credit assessment information, the above process cannot further and more accurately evaluate and analyze the information correlation between the numerous information items in the credit assessment information. Therefore, the above process cannot comprehensively and accurately assess the credit of the loan, which may cause the financial institution to lose potential customers, and may cause the customers with potential risks to obtain loans, resulting in certain loan risks for the financial institution. It can be seen that this is extremely detrimental to the development sustainability of the loan business.

[0041] Therefore, based on this, the embodiment of the present application provides a personal business loan credit assessment method based on Bayesian learning, please refer to Figure 2 , Figure 2 is a process schematic diagram of a personal business loan credit assessment method based on Bayesian learning provided by the embodiment of the present application, as Figure 2 shown, the method comprises the following steps:

[0042] 101: receiving a loan application from a target user, the loan application being used to request a guarantee loan for a target business subject.

[0043] The target business subject can be an individual industrial and commercial household, an enterprise or other forms of business subject.

[0044] The guarantee loan is a personal business loan. The personal business loan refers to a loan issued by a financial institution to a borrower for the purpose of circulating funds, purchasing or updating operating equipment, paying rental operating site rent, commercial house decoration and other legal production and operation activities of the business subject to be invested.

[0045] 102: Obtain the business plan information of the target business entity, the credit assessment information of the target user, and the historical operating condition information of other business entities under the target user.

[0046] The business plan information of the target business entity can include the planned operating scope, target customer group, strategic positioning, and the like of the target business entity.

[0047] The credit assessment information of the target user can include the personal credit information, tax information, judicial information (including information of persons subject to enforcement, legal litigation information), and black and gray list information (court black and gray list, black and gray list of the National Development and Reform Commission, and online lending black and gray list) of the target user.

[0048] In a specific implementation, before obtaining the credit assessment information of the target user, the target user needs to authorize the operation of the part of the credit assessment information that is not open to the public, and the authorization operation is used to indicate that the target user agrees to obtain the part of the credit assessment information that is not open to the public by the financial institution for processing the loan guarantee transaction.

[0049] The historical operating condition information of other business entities under the target user can include the operating scope, operating place information, and profitability of the other business entities. The other business entities can be individual industrial and commercial households, enterprises, or other forms of business entities.

[0050] The business plan information of the target business entity, the credit assessment information of the target user, and the historical operating condition information of other business entities under the target user can be included in the loan application in a specific implementation, so that the above information can be obtained in the loan application; or the above information can be obtained in the respective corresponding information channels.

[0051] 103: Analyze the operating correlation between the business plan information of the target business entity and the historical operating condition information of other business entities, and determine a first operating risk value.

[0052] The analysis of the operating correlation between the business plan information of the target business entity and the historical operating condition information of other business entities can be an analysis of the operating correlation between the business plan information of the target business entity and the historical operating condition information of other business entities based on information items of the same nature in a specific implementation.

[0053] In a specific implementation, due to the business correlation between the business plan information of the target business subject and the historical operating condition information of the other business subjects, the operating capability of the target user in the corresponding business field can be reflected, and the operating capability of the target user in the corresponding business field is an extremely important influencing factor for the operating condition of the target business subject, which greatly affects the operating development potential and profitability of the target business subject. Therefore, by analyzing the business correlation between the business plan information of the target business subject and the historical operating condition information of the other business subjects, the first operating risk value is determined, which can significantly improve the accuracy of the risk assessment of the loan application.

[0054] Further, since the same user has operating experience in a certain operating range, it will be easier to achieve success in the same operating range again. Therefore, the first operating risk value is negatively correlated with the business correlation between the business plan information of the target business subject and the historical operating condition information of the other business subjects, that is, the greater the business correlation between the business plan information of the target business subject and the historical operating condition information of the other business subjects, the lower the first operating risk value.

[0055] Exemplarily, the analysis of the business correlation between the business plan information of the target business subject and the historical operating condition information of the other business subjects is to analyze whether the operating range in the business plan information of the target business subject and the operating range in the historical operating condition information of the other business subjects have business correlation, if the operating ranges of the two are similar, it is considered that there is business correlation, and the first operating risk value of the loan application is lower.

[0056] 104: input the credit assessment information of the target user and the business plan information of the target business subject into the Bayesian learning model as input items, and obtain the second operating risk value.

[0057] Among them, Bayesian learning is to directly obtain the overall distribution by using the prior distribution of the parameter and the posterior distribution obtained from the sample information. The result of Bayesian learning is expressed as the probability distribution of a random variable, which can be understood as the degree of confidence in different possibilities.

[0058] In a specific implementation, due to the large and scattered information quantity of the credit evaluation information of the target user and the business plan information of the target business subject, the traditional risk assessment model is difficult to mine the joint influence of the specific information on the loan application risk, therefore, in order to establish the information correlation mechanism between the user personal credit and the business plan implementation to accurately assess the loan application risk, the embodiment of the application analyzes and processes the correlation between the credit evaluation information of the target user and the business plan information of the target business subject through the Bayesian learning model, and then can extract the key information affecting the loan application risk from the large and scattered information quantity, and ensure the reliability of the loan application risk assessment.

[0059] In a specific implementation, the Bayesian learning model includes a Bayesian learning formula, and the form of the Bayesian learning formula can be: P(X|A, B, C, D…) = P(X|A)*P(X|B)*P(X|C)*P(X|D)…, wherein X is an event to be occurred, A, B, C, D… are conditional events that have occurred, and P(X|A) represents a probability value of the event X occurring under the condition that the A conditional event has occurred.

[0060] Exemplarily, the loan application risk value is X, the information entries in the credit evaluation information of the target user are A, B, and C, and the information entries in the business plan information of the target business subject are A', B', and C', wherein A and A', B and B', and C and C' have information correlation, therefore, the loan application risk value X is taken as an event to be occurred, and A and A', B and B', and C and C' are taken as conditional events that have occurred, so that the Bayesian learning formula followed by the Bayesian learning model when obtaining the second business risk value is: second business risk value = P(X|A, A')*P(X|B, B')*P(X|C, C'), wherein P(X|A, A') represents the loan application risk value X under the condition that the A and A' conditional events have occurred at the same time.

[0061] The Bayesian learning model can include an input layer, a formula layer, and an output layer.

[0062] Based on the above exemplary embodiment, please refer to Figure 3 , Figure 3 is a structural diagram of a Bayesian learning model provided by the embodiment of the application, as Figure 3 shown, the Bayesian learning model includes the following levels:

[0063] The input layer is used to receive the credit evaluation information of the target user and the business plan information of the target business subject, and correspondingly substitutes the information entries A, B, and C in the credit evaluation information of the target user and the information entries A', B', and C' in the business plan information of the target business subject into the parameters in the Bayesian learning formula included in the formula layer.

[0064] a formula layer including a Bayesian learning formula: second business risk value = P(X|A, A') * P(X|B, B') * P(X|C, C'), used for calculating the second business risk value according to the information items A, B, C, A', B', C' input by the input layer and the Bayesian learning formula.

[0065] an output layer used for outputting the second business risk value calculated by the formula layer.

[0066] It should be noted that the Bayesian learning model shown in Figure 3 is only an example of a Bayesian learning model, and in specific applications, the Bayesian learning model can also exist in the form of other levels.

[0067] 105: determining an evaluation result of the loan application according to the first business risk value and the second business risk value, the evaluation result of the loan application including passing the loan application or not passing the loan application.

[0068] In the specific implementation, the evaluation result of the loan application can be that both the first business risk value and the second business risk value are less than a preset business risk value, or the sum of the first business risk value and the second business risk value is less than the preset business risk value, or the product between the first business risk value and the second business risk value is less than the preset business risk value.

[0069] Exemplarily, please refer to Figure 4 , Figure 4 is a schematic diagram of an architecture of a personal business loan credit evaluation system based on Bayesian learning provided by an embodiment of the present application, as shown in Figure 4As shown, the system comprises a server and a Bayesian learning model. A target user applies for a personal business loan from a financial institution for the business activities of a target business subject. Therefore, the target user uses a target user terminal to submit a loan application to a financial institution terminal. After the financial institution terminal obtains the business plan information of the target business subject, the credit assessment information of the target user, and the historical business status information of other business subjects under the target user, the financial institution terminal sends the above information to the personal business loan credit assessment system based on Bayesian learning. The server in the system receives the above information and analyzes the business correlation between the business plan information of the target business subject and the historical business status information of other business subjects, determines a first business risk value. At the same time, the server inputs the credit assessment information of the target user and the business plan information of the target business subject into the Bayesian learning model in the system, and the Bayesian learning model outputs a second business risk value to the server. Since the first business risk value and the second business risk value are both less than a preset business risk value, the server determines that the evaluation result of the loan application is that the target user passes the loan application, and feeds back the evaluation result to the financial institution terminal.

[0070] It can be seen that, in the embodiment of the application, the business correlation between the business plan information of the target business subject and the historical business status information of other business subjects is analyzed to determine a first business risk value. The credit assessment information of the target user and the business plan information of the target business subject are input into the Bayesian learning model to obtain a second business risk value. The evaluation result of the loan application is determined according to the first business risk value and the second business risk value. When facing a large amount of information about users and business subjects, the financial institution can comprehensively and accurately perform credit assessment on personal business loans, thereby avoiding loan risks and ensuring the sustainable development of loan business.

[0071] In one possible example, the above method further comprises:

[0072] If the first business risk value and the second business risk value are both lower than the preset business risk value, it is determined that the evaluation result of the loan application is that the target user passes the loan application.

[0073] The value range of the first business risk value and the second business risk value can be 0-1, and the preset business risk value can be 0.6, 0.8, or other business risk values.

[0074] It can be seen that, in the embodiment of the application, the evaluation result of the loan application is determined to be that the target user passes the loan application only when the first business risk value and the second business risk value are both lower than the preset business risk value. Therefore, the default risk of the target user who can obtain the loan is low, thereby avoiding loan risks and ensuring the sustainable development of loan business.

[0075] In one possible example, the training process of the Bayesian learning model described above is as follows:

[0076] Obtain a training data set, the training data set including credit evaluation information of historical users and business plan information of historical business entities in a plurality of historical loan applications;

[0077] Input the training data set into an initial learning model to obtain predicted operating risk values of the plurality of historical loan applications;

[0078] Compare the predicted operating risk values of the plurality of historical loan applications with historical operating risk values corresponding to the plurality of historical loan applications, and if the error between the predicted operating risk values and the historical operating risk values is within a preset error range, determine that the initial learning model is accurate in prediction.

[0079] Determine the prediction accuracy of the initial learning model, the prediction accuracy representing the proportion of historical loan applications with accurate prediction in the training data set;

[0080] If the prediction accuracy is less than a preset accuracy, iteratively train the initial learning model, and if the prediction accuracy of the initial learning model is higher than or equal to the preset accuracy, determine that the initial learning model is trained and obtain the Bayesian learning model.

[0081] In one possible example, the method further includes:

[0082] Obtain at least one user evaluation condition and at least one business entity evaluation condition according to the credit evaluation information of the historical users and the business plan information of the historical business entities;

[0083] Obtain at least one combined condition event according to the at least one user evaluation condition and the at least one business entity evaluation condition;

[0084] Generate a conditional probability factor according to the at least one combined condition event, the occurred condition events in the conditional probability factor corresponding to obtaining an expected second risk value and / or an unexpected second risk value, the expected second risk value being a second risk value when the evaluation result of the historical loan application is passing the loan application, the unexpected second risk value being a second risk value when the evaluation result of the historical loan application is not passing the loan application, and the to-be-occurred events in the conditional probability factor corresponding to the at least one combined condition event;

[0085] Logically combine the at least one conditional probability factor to generate the initial learning model.

[0086] The value range of the predicted operating risk value and the historical operating risk value can be 0-1, and the preset error range can be ±0.05.

[0087] The initial learning model is iteratively trained, and in a specific implementation, the historical user credit evaluation information and the specific information entries of the historical business entity business plan information included in the training data set of the input initial learning model can be adjusted, or the specific evaluation conditions included in at least one combination condition event in the conditional probability factor can be adjusted.

[0088] The preset accuracy can be in the form of a value of 0-1 or in the form of a percentage of 0-100%. If the value of 0-1 is used, the preset accuracy can be 0.8, 0.9 or other accuracy; if the percentage of 0-100% is used, the preset accuracy can be 80%, 90% or other accuracy.

[0089] The user evaluation condition is a specific information entry in the historical user credit evaluation information, and the business entity evaluation condition is a specific information entry in the historical business entity business plan information.

[0090] For example, refer to Figure 5 , Figure 5 is an example of a personal business loan credit evaluation method based on Bayesian learning provided by the embodiments of the present application, as shown in Figure 5 To generate an initial learning model, first, user evaluation conditions 1-N and business entity evaluation conditions 1-N are obtained according to historical user credit evaluation information and historical business entity business plan information, respectively; then the user evaluation conditions 1-N and the business entity evaluation conditions 1-N are combined respectively to obtain combination condition events 1-N, wherein combination condition event 1 indicates that user evaluation condition 1 and business entity evaluation condition 1 occur at the same time, and combination condition event N indicates that user evaluation condition N and business entity evaluation condition N occur at the same time; then conditional probability factors 1-N are generated according to the combination condition events 1-N, and the to-occur events of the conditional probability factors 1-N are all expected second risk values, wherein the conditional probability factor 1 can be expressed as P-1 (expected second risk value | combination condition event 1) = P-1 (expected second risk value | user evaluation condition 1, business entity evaluation condition 1), and the conditional probability factor N can be expressed as P-N (expected second risk value | combination condition event N) = P-N (expected second risk value | user evaluation condition N, business entity evaluation condition N); finally, the conditional probability factors 1-N are logically combined, and the product of the conditional probability factors 1-N is taken as the second business risk value, that is, the second business risk value = conditional probability factor 1 *... * conditional probability factor N = P-1 (expected second risk value | user evaluation condition 1, business entity evaluation condition 1) *... * P-N (expected second risk value | combination condition event N), thus completing the initial learning model generation process.

[0091] It should be noted that the initial learning model generation process as shown in Figure 5 is only an example of an initial learning model generation process, and in specific applications, the initial learning model generation process can also be implemented in other ways.

[0092] It can be seen that the training process of the Bayesian learning model provided in the embodiments of the present application, wherein the initial learning model is obtained by obtaining at least one user evaluation condition and at least one business entity evaluation condition according to the credit evaluation information of the historical user and the business plan information of the historical business entity, then obtaining at least one combined condition event according to the at least one user evaluation condition and the at least one business entity evaluation condition, then generating a conditional probability factor according to the at least one combined condition event, and finally logically combining the at least one conditional probability factor to produce, and only when the prediction accuracy of the initial learning model is higher than or equal to the preset accuracy, it is determined that the initial learning model training is successful and the Bayesian learning model is obtained. Further, using the Bayesian learning model trained by the training process provided in the embodiments of the present application to predict the second business risk value of the personal business loan application can not only liberate the personnel labor of the financial institutions, but also improve the prediction accuracy of the financial institutions for loan risk.

[0093] In one possible example, the user evaluation condition includes a first condition, the first condition representing that the tax completion rate of the historical user is greater than a preset completion rate, and the business entity evaluation condition includes a second condition, the second condition representing that the user responsibility proportion of the historical user in the business entity is greater than a preset proportion, and the at least one combined condition event is obtained according to the at least one user evaluation condition and the at least one business entity evaluation condition, including:

[0094] obtaining a combined condition event according to the occurrence of the first condition and the occurrence of the second condition; and / or obtaining a combined condition event according to the non-occurrence of the first condition and the non-occurrence of the second condition.

[0095] The tax completion rate of the historical user refers to the completion of the tax declaration of the historical user.

[0096] In specific implementation, since the tax completion rate of the historical user is better, it means that the credit ability of the historical user is good, and when the user responsibility proportion of the historical user in the business entity is larger, the participation of the historical user in the business decision of the business entity will also be higher, and the good credit ability of the historical user can be better played, that is, when the tax completion rate of the historical user is greater than the preset completion rate and the user responsibility proportion of the historical user in the business entity is greater than the preset proportion, the default risk of the loan application is relatively low, and vice versa. Therefore, a combined condition event is obtained according to the occurrence of the first condition and the occurrence of the second condition; and / or a combined condition event is obtained according to the non-occurrence of the first condition and the non-occurrence of the second condition.

[0097] The preset completion rate can be represented in the form of a value from 0 to 1 or in the form of a percentage from 0 to 100%. If represented in the form of a value from 0 to 1, the preset completion rate can be 0.8, 0.9 or other accuracy; if represented in the form of a percentage from 0 to 100%, the preset completion rate can be 80%, 90% or other accuracy.

[0098] The user responsibility proportion of the historical user in the business subject can refer to the shareholding proportion of the historical user in the business subject, or can only refer to the registered capital investment proportion of the historical user in the business subject.

[0099] Exemplarily, a combined condition event is obtained according to the first condition occurrence and the second condition occurrence, and a conditional probability factor is generated according to the combined condition event, and the conditional probability factor = (expected second risk value | first condition occurrence, second condition occurrence) = (expected second risk value | tax completion rate of historical user is greater than preset completion rate, user responsibility proportion of historical user in business subject is greater than preset proportion).

[0100] Based on the above exemplary embodiments, in specific implementation, in the case that the tax completion rate of the historical user is greater than the preset completion rate and the user responsibility proportion of the historical user in the business subject is greater than the preset proportion, the corresponding second operating risk value can be set according to the second operating risk value corresponding to the historical loan application in which the same combined condition event occurs, for example, the average second operating risk value corresponding to all loan applications in which the tax completion rate of the historical user is greater than the preset completion rate and the user responsibility proportion of the historical user in the business subject is greater than the preset proportion is 0.2, and the second operating risk value corresponding to the combined condition event is determined as 0.2.

[0101] It can be seen that, in the embodiments of the present application, for the first condition and the second condition having certain information correlation or having certain causal relationship, a combined condition event is obtained according to the occurrence of the first condition and the occurrence of the second condition, and / or a combined condition event is obtained according to the non-occurrence of the first condition and the non-occurrence of the second condition, so that the combined condition events used by the Bayesian learning model in the training process have a common decision effect on the second operating risk value, and then the Bayesian learning model trained by the training process provided by the embodiments of the present application can comprehensively and accurately predict the second operating risk value of the personal operating loan application, and improve the prediction accuracy of the financial institutions for the loan risk.

[0102] It should be noted that in the above embodiment, the first condition represents that the tax completion rate of the historical user is greater than the preset completion rate, and the second condition represents that the user responsibility ratio of the historical user in the business subject is greater than the preset ratio, which is only an example of the content of the first condition and the second condition. In specific implementation, as long as the first condition and the second condition meet the requirements of having certain information correlation or having certain causal relationship, that is, the content of the first condition and the second condition is not limited here, and the first condition and the second condition can also exist in the form of representing other contents.

[0103] Based on this, in one possible example, the user evaluation condition includes a first condition, the first condition represents that the number of intellectual property rights under the historical user is greater than a preset number, and the business subject evaluation condition includes a second condition, the second condition represents that the R&D expense ratio of the business subject is greater than a preset ratio. According to at least one user evaluation condition and at least one business subject evaluation condition, at least one combined condition event is obtained, including:

[0104] According to the occurrence of the first condition and the occurrence of the second condition, a combined condition event is obtained; and / or according to the non-occurrence of the first condition and the non-occurrence of the second condition, a combined condition event is obtained.

[0105] Among them, the intellectual property rights can include at least one of the following: trademark rights, copyright, patent rights.

[0106] Among them, the R&D expense ratio refers to the ratio of R&D expenses to sales revenue. The higher the R&D expense ratio, the higher the technological content of the products of the business subject, which is a technology-based enterprise.

[0107] In one possible example, the business plan information of the target business subject includes the business scope of the target business subject, and the historical business condition information includes the business scope of the other business subjects. The business correlation between the business plan information of the target business subject and the historical business condition information of the other business subjects is analyzed to determine a first business risk value, including:

[0108] The business scope of the target business subject and the business scope of the other business subjects are analyzed to determine the business scope correlation between the business scope of the target business subject and the business scope of the other business subjects. The business scope correlation includes different industries, the same industry and the same coverage, and the same industry and different coverage.

[0109] According to the business scope correlation, the business correlation index between the business plan information and the historical business condition information is determined;

[0110] The first business risk value is determined according to the business correlation index. The business correlation index and the first business risk value are negatively correlated.

[0111] In one possible example, the business plan information of the target business entity further includes business site information of the target business entity, and the historical business status information further includes business site information of the other business entities. The method further includes:

[0112] If the business scope correlation relationship is the same industry with the same coverage or the same industry with different coverage, the distance between the business site information of the target business entity and the business site information of the other business entities is analyzed to determine the business site distance between the target business entity and the other business entities.

[0113] The business correlation index between the business plan information and the historical business status information is determined according to the business scope correlation relationship, including:

[0114] The first business risk value of the target business entity is determined according to the business scope correlation relationship and the business site distance, wherein the first business risk value is positively correlated with the business site distance when the business scope correlation relationship is the same industry with different coverage, and the first business risk value is negatively correlated with the business site distance when the business scope correlation relationship is the same industry with the same coverage.

[0115] The business scope refers to the production and service projects that the business entity can engage in.

[0116] The business scope correlation relationship between the business scope of the target business entity and the business scope of the other business entities is determined by analyzing the business scope of the target business entity and the business scope of the other business entities. In specific implementation, the National Economic Industry Classification can be used as a basis.

[0117] For example, if the business scope of the target business entity is “food and beverage specialty retail”, and the business scope of the other business entity is “household appliances and electronic products specialty retail”, the business scope correlation relationship between the business scope of the target business entity and the business scope of the other business entity is determined to be different industries.

[0118] For another example, if the business scope of the target business entity and the other business entity is “beverage and tea retail” under “food and beverage specialty retail”, the business scope correlation relationship between the business scope of the target business entity and the business scope of the other business entity is determined to be the same industry with the same coverage.

[0119] For another example, if the business scope of the target business entity is “beverage and tea retail” under “food and beverage specialty retail”, and the business scope of the other business entity is “cake and bread retail” under “food and beverage specialty retail”, the business scope correlation relationship between the business scope of the target business entity and the business scope of the other business entity is determined to be the same industry with different coverage.

[0120] In the specific implementation, the business correlation index between the business plan information and the historical business status information can be determined according to the business scope correlation relationship. In the specific implementation, the business correlation index when the business scope correlation relationship is different industries is less than that when the business scope correlation relationship is the same industry and different coverage ranges or the same industry and the same coverage range, so that the first business risk value when the business scope correlation relationship is different industries is greater than that when the business scope correlation relationship is the same industry and different coverage ranges or the same industry and the same coverage range.

[0121] In the specific implementation, according to the agglomeration economic effect, when a consumer consumes a certain commodity at a certain location, the consumer will also consume complementary or auxiliary products of the commodity. For example, if a consumer purchases a cake at a cake shop, the consumer will also purchase beverages at a nearby beverage shop to eat with the cake. Therefore, the first business risk value and the business location distance are in a positive correlation relationship when the business scope correlation relationship is the same industry and different coverage ranges.

[0122] In the specific implementation, because the commodities in the same industry and the same coverage range belong to substitutes, most consumers will only choose one to consume. That is, the commodities in the same industry and the same coverage range are in a competitive relationship at the same location or nearby locations. Therefore, the first business risk value and the business location distance are in a negative correlation relationship when the business scope correlation relationship is the same industry and the same coverage range.

[0123] It can be seen that, in the embodiment of the application, the business plan information of the target business subject includes the business scope and the business site information of the target business subject, the historical business status information includes the business scope and the business site information of other business subjects, by analyzing the business scope of the target business subject and the business scope of other business subjects, the business scope correlation between the business scope of the target business subject and the business scope of other business subjects is determined, then according to the business scope correlation, the business correlation index between the business plan information and the historical business status information is determined, and finally according to the business correlation index, the first business risk value is determined. Therefore, when the business scope correlation is different industries, the first business risk value is higher, and when the business scope correlation is the same industry, the first business risk value is determined according to the further same industry coverage range and the distance between the business sites. Since the business correlation index and the first business risk value are negatively correlated, specifically, when the business scope correlation is the same industry and the same coverage range, the closer the distance between the business sites of the target business subject and other business subjects, the greater the first business risk value, and when the business scope correlation is the same industry and different coverage ranges, the closer the distance between the business sites of the target business subject and other business subjects, the smaller the first business risk value. By using the method of the embodiment of the application, the financial institution can accurately analyze the business correlation between the business plan information of the target business subject and the historical business status information of other business subjects, thereby accurately performing credit assessment on personal business loans, and further avoiding loan risks and ensuring the sustainable development of loan business.

[0124] Consistent with the above Figure 2 indicated embodiment, please refer to Figure 6 , Figure 6 A structure schematic diagram of a personal business loan credit assessment device based on Bayesian learning provided by the embodiment of the application is shown in Figure 6 .

[0125] A personal business loan credit assessment device based on Bayesian learning, the device comprises:

[0126] 201: a receiving unit, configured to receive a loan application from a target user, the loan application being used to request a target business subject to obtain a guaranteed loan.

[0127] 202: an acquisition unit, configured to acquire business plan information of the target business subject, credit assessment information of the target user, and historical business status information of other business subjects under the target user.

[0128] 203: an analysis unit, configured to analyze the business correlation between the business plan information of the target business subject and the historical business status information of other business subjects, and determine a first business risk value.

[0129] 204: A model unit, configured to input the target user's credit assessment information and the target business entity's business plan information as input items into a Bayesian learning model to obtain a second business risk value.

[0130] 205: An evaluation unit, configured to determine an evaluation result of the loan application based on the first business risk value and the second business risk value, wherein the evaluation result of the loan application includes approval or rejection of the loan application.

[0131] It can be seen that in the device provided by the embodiment of the present application, the business correlation between the business plan information of the target business entity and the historical business status information of other business entities is analyzed by the analysis unit to determine the first business risk value, and the credit assessment information of the target user and the business plan information of the target business entity are input as input items into the Bayesian learning model through the model unit to obtain the second business risk value, and then the evaluation unit determines the evaluation result of the loan application based on the first business risk value and the second business risk value. Using the device of the embodiment of the present application, when faced with a large amount of multifaceted information from users and business entities, financial institutions can also conduct comprehensive and accurate credit assessments for personal business loans, thereby avoiding loan risks and ensuring the sustainable development of the loan business.

[0132] Specifically, the embodiment of the present application can divide the functional units of the personal business loan credit assessment device based on Bayesian learning according to the above method example. For example, each functional unit can be divided corresponding to each function, or two or more functions can be integrated into one processing unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. It should be noted that the division of units in the embodiment of the present application is schematic and is only a logical functional division. There may be other division methods in actual implementation.

[0133] With the above Figure 2 In accordance with the embodiment shown, the present application embodiment provides an electronic device, see Figure 7 , Figure 7 This is a server structure diagram of a hardware operating environment of an electronic device provided by an embodiment of the present application, such as Figure 7 As shown, the electronic device includes a processor, a memory, and computer execution instructions stored in the memory and executable on the processor. When the computer execution instructions are executed, the electronic device executes instructions including any step of the personal business loan credit assessment method based on Bayesian learning.

[0134] Among them, the processor is CPU (Central Processing Unit).

[0135] The memory, which can be a high-speed RAM memory or a stable memory such as a disk memory, is optional.

[0136] Those skilled in the art can understand that Figure 7 The structure of the server shown in the above embodiments is not limited thereto, and can include more or fewer components, or combine certain components, or different component arrangements.

[0137] As shown in the above embodiments, Figure 7 The memory can include an operating system, a network communication module, and computer execution instructions of the personal business loan credit assessment method based on Bayesian learning. The operating system is used to manage and control the hardware and software resources of the server, and supports the running of the computer execution instructions. The network communication module is used to realize the communication between the components in the memory, and the communication with other hardware and software in the server. The communication can use any communication standard or protocol, including but not limited to GSM (Global System of Mobile communication), GPRS (General Packet Radio Service), CDMA2000 (Code Division Multiple Access 2000), WCDMA (Wideband Code Division Multiple Access), TD-SCDMA (Time Division-Synchronous Code Division Multiple Access), etc.

[0138] In the server shown in the above embodiments, Figure 7 In the server shown in the above embodiments, the processor is used to execute the computer execution instructions of the personnel management stored in the memory, and realize the following steps: receiving a loan application from a target user, the loan application being used to request a guaranteed loan for the target business subject; obtaining business plan information of the target business subject, credit assessment information of the target user, and historical business status information of other business subjects under the target user; analyzing the business correlation between the business plan information of the target business subject and the historical business status information of the other business subjects, and determining a first business risk value; inputting the credit assessment information of the target user and the business plan information of the target business subject into a Bayesian learning model as input items, and obtaining a second business risk value; and determining an evaluation result of the loan application according to the first business risk value and the second business risk value, the evaluation result of the loan application including passing the loan application or not passing the loan application.

[0139] The specific implementation of the server involved in the present application can be referred to the above embodiments of the personal business loan credit evaluation method based on Bayesian learning, and will not be repeated here.

[0140] The embodiment of the present application provides a computer readable storage medium, and the computer readable storage medium stores computer instructions. When the computer instructions run on a communication device, the communication device performs the following steps: receiving a loan application from a target user, the loan application being used to request a guaranteed loan for a target business subject; obtaining business plan information of the target business subject, credit evaluation information of the target user, and historical operation condition information of other business subjects under the target user; analyzing operation correlation between the business plan information of the target business subject and the historical operation condition information of the other business subjects, and determining a first operation risk value; inputting the credit evaluation information of the target user and the business plan information of the target business subject into a Bayesian learning model as input items, and obtaining a second operation risk value; and determining an evaluation result of the loan application according to the first operation risk value and the second operation risk value, the evaluation result of the loan application including passing the loan application or not passing the loan application. The computer includes an electronic device.

[0141] The electronic terminal device includes a mobile phone, a tablet computer, a personal digital assistant, a wearable device, and the like.

[0142] The computer readable storage medium can be an internal storage unit of the electronic device, for example, a hard disk or a memory of the electronic device. The computer readable storage medium can also be an external storage device of the electronic device, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like. Further, the computer readable storage medium can include both the internal storage unit and the external storage device of the electronic device. The computer readable storage medium is used to store computer execution instructions and other computer execution instructions and data required by the electronic device. The computer readable storage medium can also be used to temporarily store data that has been output or will be output.

[0143] The specific implementation of the computer readable storage medium involved in the present application can be referred to the above embodiments of the personal business loan credit evaluation method based on Bayesian learning, and will not be repeated here.

[0144] The embodiment of the present application provides a computer program product, wherein the computer program product includes a computer program, and the computer program is operable to make the computer perform part or all steps of any one of the personal business loan credit evaluation methods based on Bayesian learning described in the above method embodiments. The computer program product can be a software installation package.

[0145] It should be noted that for the foregoing embodiments of the personal business loan credit assessment method based on Bayesian learning, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the order of the described actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions involved are not necessarily necessary for the present application.

[0146] The embodiments of the present application are described in detail above, and in this paper, specific examples are applied to describe the principles and implementation methods of the personal business loan credit assessment method based on Bayesian learning and related products of the present application. The above embodiment description is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the personal business loan credit assessment method based on Bayesian learning and related products of the present application, the specific implementation method and application range will be changed; in summary, the content of the specification should not be understood as a limitation of the present application.

[0147] The present application is described with reference to the flowcharts and / or block diagrams of the method, hardware product and computer program product of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of the flows and / or blocks in the flowchart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a machine for implementing the functions specified in the flowchart and / or block diagram. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The device for realizing the functions specified in one block or multiple blocks.

[0148] Although the present application is described herein in conjunction with various embodiments, other variations of the disclosed embodiments can be understood and implemented by those skilled in the art with reference to the attached drawings, disclosure, and appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "one" does not exclude a plurality. Certain measures described in mutually different dependent claims can be combined and produce desirable results.

[0149] Obviously, those skilled in the art can make various modifications and variations to the personal business loan credit assessment method based on Bayesian learning and related products provided in the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A personal business loan credit assessment method based on Bayesian learning, characterized by: The method comprises: receiving a loan application from a target user, wherein the loan application is used to request a secured loan for a target business entity; Obtaining business plan information of the target business entity, credit assessment information of the target user, and historical operating status information of other business entities under the name of the target user; Analyzing the business relevance between the business plan information of the target business entity and the historical business status information of the other business entities to determine a first business risk value; Inputting the target user's credit assessment information and the target business entity's business plan information into the Bayesian learning model as input items to obtain a second business risk value; determining an evaluation result of the loan application based on the first operating risk value and the second operating risk value, wherein the evaluation result of the loan application includes approving the loan application or rejecting the loan application; The training process of the Bayesian learning model includes: obtaining a training data set, wherein the training data set includes credit assessment information of historical users and business plan information of historical business entities in multiple historical loan applications; the method further includes: At least one user evaluation condition and at least one business entity evaluation condition are obtained based on the credit evaluation information of the historical user and the business plan information of the historical business entity; wherein the user evaluation condition includes a first condition, wherein the first condition indicates that the tax payment completion rate of the historical user is greater than a preset completion rate, and the business entity evaluation condition includes a second condition, wherein the second condition indicates that the user responsibility ratio of the historical user in the business entity is greater than a preset ratio; At least one combined condition event is obtained according to the at least one user evaluation condition and the at least one business entity evaluation condition, including: The combined conditional event is obtained based on the occurrence of the first condition and the occurrence of the second condition; and / or the combined conditional event is obtained based on the non-occurrence of the first condition and the non-occurrence of the second condition; A conditional probability factor is generated based on the at least one combined conditional event, the conditional event that has occurred in the conditional probability factor corresponds to an expected second risk value and / or an unexpected second risk value, the expected second risk value is the second risk value when the evaluation result of the historical loan application is that the loan application is approved, and the unexpected second risk value is the second risk value when the evaluation result of the historical loan application is that the loan application is rejected, and the event to occur in the conditional probability factor corresponds to the at least one combined conditional event; the at least one conditional probability factor is logically combined to generate an initial learning model, and the initial learning model is used to train to obtain a Bayesian learning model.

2. The method according to claim 1, characterized in that The method further comprises: If the first operating risk value and the second operating risk value are both lower than a preset operating risk value, the evaluation result of the loan application is determined to be the approval of the loan application.

3. The method according to claim 1 or 2, characterized in that The training process of the Bayesian learning model also includes: Inputting the training data set into the initial learning model to obtain predicted operating risk values ​​of multiple historical loan applications; Comparing the predicted operating risk values ​​of the multiple historical loan applications with the historical operating risk values ​​corresponding to the multiple historical loan applications, and determining that the prediction of the initial learning model is accurate if the error between the predicted operating risk value and the historical operating risk value is within a preset error range; Determining a prediction accuracy of the initial learning model, wherein the prediction accuracy represents a proportion of accurately predicted historical loan applications in the training data set; When the prediction accuracy is less than the preset accuracy, the initial learning model is iteratively trained. If the prediction accuracy of the initial learning model is higher than or equal to the preset accuracy, it is determined that the training of the initial learning model is completed and the Bayesian learning model is obtained.

4. The method according to claim 1, wherein The business plan information of the target business entity includes the business scope of the target business entity, and the historical business status information includes the business scope of the other business entities. The analyzing the business correlation between the business plan information of the target business entity and the historical business status information of the other business entities to determine the first business risk value includes: Analyze the business scope of the target business entity and the business scope of the other business entities to determine the business scope correlation relationship between the business scope of the target business entity and the business scope of the other business entities, where the business scope correlation relationship includes different industries, the same scope of coverage in the same industry, and different scopes of coverage in the same industry; Determining a business correlation index between the business plan information and the historical business status information based on the business scope correlation relationship; The first operating risk value is determined according to the operating correlation index, and there is a negative correlation between the operating correlation index and the first operating risk value.

5. The method according to claim 4, characterized in that The business plan information of the target business entity also includes the business location information of the target business entity, and the historical business status information also includes the business location information of the other business entities. The method further includes: If the business scope association relationship is the same coverage of the same industry or different coverage of the same industry, then analyzing the distance between the business location information of the target business entity and the business location information of the other business entities to determine the business location distance between the target business entity and the other business entities; The determining of the business correlation index between the business plan information and the historical business status information based on the business scope correlation relationship includes: The first operating risk value of the target business entity is determined based on the business scope association relationship and the distance to the business premises, wherein, when the business scope association relationship is the same industry with different coverage scopes, the first operating risk value is positively correlated with the distance to the business premises, and when the business scope association relationship is the same industry with the same coverage scope, the first operating risk value is negatively correlated with the distance to the business premises.

6. A personal business loan credit assessment device based on Bayesian learning, characterized in that: The device comprises: a receiving unit, configured to receive a loan application from a target user, wherein the loan application is used to request a secured loan for a target business entity; an acquisition unit, configured to acquire business plan information of the target business entity, credit assessment information of the target user, and historical operating status information of other business entities under the target user; an analyzing unit, configured to analyze the business relevance between the business plan information of the target business entity and the historical business status information of the other business entities, and determine a first business risk value; a model unit, configured to input the credit assessment information of the target user and the business plan information of the target business entity as input items into the Bayesian learning model to obtain a second business risk value; an evaluation unit, configured to determine an evaluation result of the loan application based on the first operating risk value and the second operating risk value, wherein the evaluation result of the loan application includes approving the loan application or rejecting the loan application; The training process of the Bayesian learning model includes: obtaining a training data set, wherein the training data set includes credit assessment information of historical users and business plan information of historical business entities in multiple historical loan applications; the device is further used to: At least one user evaluation condition and at least one business entity evaluation condition are obtained based on the credit evaluation information of the historical user and the business plan information of the historical business entity; wherein the user evaluation condition includes a first condition, wherein the first condition indicates that the tax payment completion rate of the historical user is greater than a preset completion rate, and the business entity evaluation condition includes a second condition, wherein the second condition indicates that the user responsibility ratio of the historical user in the business entity is greater than a preset ratio; At least one combined condition event is obtained according to the at least one user evaluation condition and the at least one business entity evaluation condition, including: The combined conditional event is obtained based on the occurrence of the first condition and the occurrence of the second condition; and / or the combined conditional event is obtained based on the non-occurrence of the first condition and the non-occurrence of the second condition; generating a conditional probability factor based on the at least one combined conditional event, wherein the conditional event that has occurred in the conditional probability factor corresponds to obtaining an expected second risk value and / or an unexpected second risk value, the expected second risk value being the second risk value when the evaluation result of the historical loan application is an approval of the loan application, and the unexpected second risk value being the second risk value when the evaluation result of the historical loan application is a rejection of the loan application, and the event to occur in the conditional probability factor corresponding to the at least one combined conditional event; The at least one conditional probability factor is logically combined to generate an initial learning model, where the initial learning model is used to train a Bayesian learning model.

7. An electronic device, characterized in that: The electronic device comprises a processor, a memory, and computer-executable instructions stored in the memory and executable on the processor. When the computer-executable instructions are executed, the electronic device executes the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and when the computer instructions are executed on the communication device, the communication device is caused to execute the method according to any one of claims 1 to 5.

Images