CN115187393A

CN115187393A - Loan risk detection method and device, electronic equipment and readable storage medium

Info

Publication number: CN115187393A
Application number: CN202211113472.9A
Authority: CN
Inventors: 陈涛涛
Original assignee: Shenzhen Mingyuan Cloud Technology Co Ltd
Current assignee: Shenzhen Mingyuan Cloud Technology Co Ltd
Priority date: 2022-09-14
Filing date: 2022-09-14
Publication date: 2022-10-14

Abstract

The application discloses a loan risk detection method, a loan risk detection device, electronic equipment and a readable storage medium, wherein the loan risk detection method comprises the following steps: acquiring user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level; performing loan risk prediction at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results at the different loan risk levels; and detecting whether the target user has loan risks or not according to the loan risk prediction results. The method and the device solve the technical problem that in the prior art, the accuracy and the interpretability of loan risk prediction cannot be considered at the same time.

Description

Loan risk detection method and device, electronic equipment and readable storage medium

Technical Field

The application belongs to the technical field of financial science and technology, and relates to a loan risk detection method and device, electronic equipment and a readable storage medium.

Background

The loan risk refers to the condition that a borrower cannot pay for a loan, the house loan is longer than other loan business time periods, and more accurate prediction needs to be carried out on the risk before the loan, so that the borrower with high risk is refused.

The conventional loan risk prediction method is divided into two methods, one method is to carry out loan risk prediction based on a traditional machine learning model and is difficult to capture the correlation among different sample characteristics so as to influence the accuracy of loan risk prediction, and the other method is to carry out loan risk prediction based on a deep learning model and is high in accuracy.

Disclosure of Invention

The application mainly aims to provide a loan risk detection method, and aims to solve the technical problem that in the prior art, the accuracy and the interpretability of loan risk prediction cannot be considered at the same time.

In order to achieve the above object, the present application provides a loan risk detection method, including:

acquiring user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level;

performing loan risk prediction at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results at the different loan risk levels;

and detecting whether the target user has loan risks or not according to the loan risk prediction results.

Optionally, the loan risk levels include an average risk level, the user data includes macro economic index data, the loan risk prediction model includes an average loan risk prediction model corresponding to the average risk, and the step of performing loan risk prediction in different loan risk levels by using the user data and the loan risk prediction models corresponding to different loan risk levels includes:

constructing an average risk influence characteristic corresponding to the average risk according to the macro economic index data, wherein the macro economic index data at least comprises one of national GDP data, loan person city GDP data and unemployment rate information;

and performing loan risk prediction on the target user under the average risk hierarchy by inputting the average risk influence characteristics into the average loan risk prediction model.

Optionally, the loan risk level includes a professional industry risk level, the user data includes industry-related data, the loan risk prediction model includes a professional industry loan risk prediction model corresponding to the professional industry risk level, and the step of performing loan risk prediction at different loan risk levels by using the user data and the loan risk prediction models corresponding to different loan risk levels includes:

according to the industry related data, constructing a practitioner industry risk level influence characteristic corresponding to the practitioner industry risk level, wherein the industry related data at least comprises one of industry average income, industry average employment duration and the number of enterprises in the industry;

and performing loan risk prediction on the target user under the professional industry risk level by inputting the professional industry risk level influence characteristics into the professional industry loan risk prediction model.

Optionally, the loan risk levels include individual risk levels, the user data includes at least one of personal credit, personal income, and age, and the step of performing loan risk prediction at different loan risk levels respectively using the user data and a loan risk prediction model corresponding to the different loan risk levels includes:

constructing individual risk influence characteristics corresponding to the individual risk levels according to the individual credit, the individual income and the age;

and performing loan risk prediction on the target user under the individual risk level by inputting the individual risk influence characteristics into an individual loan risk prediction model.

Optionally, the loan risk prediction result includes a loan risk assessment value, and the step of detecting whether the target user has a loan risk according to each loan risk prediction result includes:

carrying out weighted summation on each loan risk assessment value to obtain a total loan risk assessment value;

if the total loan risk assessment value is smaller than a preset risk threshold value, judging that the target user does not have loan risk;

and if the total loan risk assessment value is not less than a preset risk threshold value, judging that the target user has loan risk.

Optionally, before the step of determining that the target user does not have a loan risk if the total borrower risk assessment value is smaller than a preset risk threshold, the loan risk detection method further includes:

acquiring the loan risk information and income expectation of at least one loaned user;

and setting the preset risk threshold according to the risk information of each borrower and the income expectation.

Optionally, after the step of determining that the target user has a loan risk if the total borrower risk assessment value is not less than a preset risk threshold, the method further includes:

obtaining a weighted summation weight corresponding to the loan risk assessment value;

and determining factors influencing the total evaluation value of the borrower risk according to the weighted sum weight and the evaluation values of different loan risk levels.

The application also provides a loan risk detection device, loan risk detection device includes:

the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring user data of a target user in different loan risk levels, and the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level;

the risk prediction module is used for performing loan risk prediction at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results at the different loan risk levels;

and the risk detection module is used for detecting whether the target user has loan risks or not according to the loan risk prediction results.

Optionally, the loan risk hierarchy includes an average risk hierarchy, the user data includes macro economic indicator data, the loan risk prediction model includes an average loan risk prediction model corresponding to the average risk, and the risk prediction module is further configured to:

and performing loan risk prediction on the target user under the average risk level by inputting the average risk influence characteristics into the average loan risk prediction model.

Optionally, the loan risk level includes a practice industry risk level, the user data includes industry-related data, the loan risk prediction model includes a practice industry loan risk prediction model corresponding to the practice industry risk level, and the risk prediction module is further configured to:

and performing loan risk prediction on the target user under the working industry risk level by inputting the working industry risk level influence characteristics into the working industry loan risk prediction model.

Optionally, the loan risk hierarchy comprises an individual risk hierarchy, the user data comprises at least one of personal credit, personal income, and age, and the risk prediction module is further configured to:

Optionally, the loan risk prediction result includes a loan risk assessment value, and the risk detection module is further configured to:

Optionally, the loan risk detection apparatus is further configured to:

Optionally, the loan risk detection device is further configured to:

The present application further provides an electronic device, the electronic device including: a memory, a processor, and a program of the loan risk detection method stored on the memory and executable on the processor, the program of the loan risk detection method, when executed by the processor, implementing the steps of the loan risk detection method as described above.

The present application also provides a computer-readable storage medium having stored thereon a program for implementing a loan risk detection method, the program implementing the steps of the loan risk detection method as described above when executed by a processor.

The present application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the loan risk detection method as described above.

Compared with the technical means of loan risk prediction based on a traditional machine learning model or loan risk prediction based on a deep learning model in the prior art, the loan risk detection method and device based on the deep learning model firstly acquire user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of average risk, working industry risk level deviation and individual risk deviation, so that loan risk prediction is respectively performed in different loan risk levels by using the user data and the loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results of different loan risk levels, and the loan risk prediction results of the target user are respectively evaluated from the three levels of average risk, working industry risk level deviation and individual risk deviation, and further whether the target user has loan risk or not are comprehensively detected according to the loan risk prediction results. The loan risk prediction model is a deep learning model, so that the accuracy of loan risk prediction results of three levels, namely average risk, industry-of-practice risk level deviation and individual risk deviation, can be ensured, the accuracy of detecting whether the target user has loan risk can be ensured, and when the loan risk exists, the influence degree of the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation, on the final detection of whether the target user has loan risk can be explained according to the loan risk prediction results of the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation, so that the interpretability of the loan risk detection on the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation is ensured.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and, together with the description, serve to explain the principles of the application.

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.

FIG. 1 is a schematic flow chart illustrating a loan risk detection method according to a first embodiment of the application;

FIG. 2 is a schematic structural diagram of an embodiment of the loan risk detection apparatus of the present application;

fig. 3 is a schematic structural diagram of the hardware operating environment related to the loan risk detection method in the embodiment of the present application.

The objectives, features, and advantages of the present application will be further described with reference to the accompanying drawings.

Detailed Description

In order to make the aforementioned objects, features and advantages of the present application more comprehensible, embodiments accompanying figures are described in detail below. It is to be understood that the embodiments described are only a few embodiments of the present application and not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without any creative effort belong to the protection scope of the present application.

Example one

In a first embodiment of the loan risk detection method, the loan risk detection method includes:

step S10, obtaining user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level;

in this embodiment, it should be noted that the loan may be a home loan, and the different loan risk levels may be one or more of an average risk level, a professional risk level, and an individual risk level, which indicates that the different loan risk levels have different influences on the final assessment of the loan risk of what.

The influence factors of each level in the average risk level, the professional risk level and the individual risk level are different, and the weights given to the different risk levels are also different according to the current social situation, so that the different loan risk levels have different influences on the final loan risk assessment.

The average risk level refers to the average risk of all the borrowers in the whole society, for example, the possibility that all the borrowers have repayment capacity is increased when the loan company works economically; when the economy goes down, the possibility that all the borrowers have repayment capacity is reduced; the different loan risk levels and the professional risk levels refer to that the industrial risk of the target user is integrally deviated from the average risk, for example, although some industries have higher treatment during employment, the opportunities of an arbitrator are also larger, and the industries cannot be measured by personal income; for example, recently, when some industry is exposed, income is reported to be higher than actual income, so that employees can obtain house credits more easily, and therefore the risk level of the industry needs to be detected; the individual risk tier refers to the borrower's individual risk.

As an example, step S10 includes obtaining user data of the target user at different loan risk levels, which may be one of an average risk level, a professional risk level and an individual risk level,

step S20, loan risk prediction is respectively carried out at different loan risk levels by utilizing the user data and the loan risk prediction models corresponding to the different loan risk levels, so as to obtain loan risk prediction results at the different loan risk levels;

in this embodiment, it should be noted that the loan risk prediction result may be a loan risk probability, the loan risk probability may be a probability that the user cannot normally repay after making a loan, the loan risk prediction result may also be a loan risk level, and the higher the loan risk level is, the higher the probability that the user cannot normally repay after making a loan is; the user data at least comprises one of first user data corresponding to an average risk level, second user data corresponding to a working industry risk level and third user data corresponding to an individual risk level, and the loan risk prediction model at least comprises one of a first risk prediction model corresponding to the average risk level, a second risk prediction model corresponding to the working industry risk level and a third risk prediction model corresponding to the individual risk level.

As an example, step S20 includes: the probability that the target user cannot normally repay after loan is evaluated on an average risk level by inputting the first user data into the first risk prediction model, so as to obtain a first loan risk probability; and/or by inputting the second user data into the second risk prediction model, evaluating the probability that the target user cannot normally repay after loan on a professional industry risk level to obtain a second loan risk probability; and/or evaluating the probability that the target user cannot pay normally after loan on the average risk level by inputting the third user data into the third risk prediction model to obtain a third loan risk probability.

As an example, a first loan risk level is obtained by inputting the first user data into the first risk prediction model and evaluating the risk level that the target user cannot normally repay after loan on an average risk level; and/or by inputting the second user data into the second risk prediction model, evaluating the risk level that the target user cannot normally repay after loan on the industry risk level to obtain a second loan risk level; and/or evaluating the risk level of the target user who can not normally repay after loan on the average risk level by inputting the third user data into the third risk prediction model to obtain a third loan risk level;

step S30, detecting whether the target user has loan risks or not according to each loan risk prediction result;

in this embodiment, it should be noted that the loan risk prediction result may include at least one of a first loan risk probability, a second loan risk probability, and a third loan risk probability, and may also include at least one of a first loan risk level, a second loan risk level, and a third loan risk level.

As an example, step S30 includes: carrying out weighted average on the first loan risk probability, the second loan risk probability and the third loan risk probability to obtain an average probability value; if the average probability value is larger than a preset probability threshold value, judging that the target user has a loan risk; and if the average probability value is not greater than a preset probability threshold value, judging that the target user does not have loan risk.

As an example, summing the first loan risk level, the second loan risk level and the third loan risk level to obtain a comprehensive loan risk level; if the comprehensive loan risk level is larger than a preset risk level threshold, judging that the target user has a loan risk; and if the comprehensive loan risk level is not greater than a preset risk level threshold, judging that the target user does not have a loan risk.

As an example, the loan risk level at least includes one of a first risk level, a second risk level and a third risk level, the user data of the first risk of the target user includes macro economic indicator data, the user data of the second risk of the target user includes data related to the industry of working, the user data of the third risk of the target user includes data related to personal repayment capacity, and the loan risk prediction model includes a prediction model of the first risk level corresponding to the first risk level, a prediction model of the second risk level corresponding to the second risk level, and a prediction model of the third risk level corresponding to the third risk level.

Wherein the loan risk hierarchy comprises an average risk hierarchy, the user data comprises macro economic indicator data, the loan risk prediction model comprises an average loan risk prediction model corresponding to the average risk,

the loan risk prediction is performed at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels, and the loan risk prediction method comprises the following steps:

step A10, constructing an average risk influence characteristic corresponding to the average risk according to the macro economic index data, wherein the macro economic index data at least comprises one of national GDP data, GDP data of the loan people in the city and unemployment rate information;

step A20, the loan risk prediction is carried out on the target user under the average risk level by inputting the average risk influence characteristics into the average loan risk prediction model.

In this embodiment, it should be noted that the average risk influencing characteristic is a result that is composed of data in the macro economic indicator data and can influence the average risk level, for example, the average risk influencing characteristic includes at least one of national GDP data, loan person city GDP data, and unemployment rate information, where the national GDP data may be time-series data for describing distribution of GDPs across the country in recent years, for example, distribution in near 10 years, or distribution in near 5 years, and the national GDP data may describe a first GDP index parameter, such as an average value, a variance, or a standard deviation, of the national GDP variation; the city GDP data may be time series data for describing the distribution of GDPs in each region of the city in recent years, such as the distribution in the last 10 years, or the distribution in the last 5 years, and may describe a second GDP index parameter of the change of the city GDP, such as an average value, a variance, a standard deviation, or the like; the loss rate data can be a loss rate index used for describing the loss condition of each industry in the society, and the loss rate index can be the loss rate per se and can also be statistical parameters such as variance, average value, standard deviation and the like among the loss rates of different industries.

As an example, steps a10 to a20 include: acquiring national GDP data, GDP data of a city where a target user is located and the loss rate information; weighting the corresponding failure rates of different industries in the failure rate information according to preset industry weight information to obtain weighted failure rate information; constructing an average risk influence characteristic corresponding to the average risk according to the national GDP data, the urban GDP data and the weighted unemployment rate information; and performing loan risk prediction on the target user under the average risk hierarchy by inputting the average risk influence characteristics into the average loan risk prediction model.

As an example, the national GDP data may be a first feature vector composed of a plurality of first GDP index parameters, the urban GDP data may be a second feature vector composed of a plurality of first GDP index parameters, the loss rate information may be a third feature vector composed of loss rate indexes of a plurality of industries, the preset industry weight information may be a weight vector composed of weight values corresponding to different industries, the weight value is used for representing the development status of the corresponding industry, the better the development status of the industry, the lower the weight value, the worse the development status of the industry, the higher the weight value, the weight vector may be periodically updated according to external hotspot industry information, for example, if the current computer industry is a hotspot industry, the industry development status is an upward trend, the corresponding weight value may be set to 0.5, if the current computer industry is an industry, the industry development status is a downward trend, the corresponding weight value may be set to 1, if the current computer industry production industry is a non-hotspot industry, the industry development status is an upward trend, the corresponding weight value may be set to 1.5, if the industry development status is a non-hotspot industry hotspot, the industry development status is a downward trend, the corresponding weight value may be set to 1.5, and the industry development status may be set to a corresponding hotspot, and the step may further include:

weighting the loss rate index of the corresponding industry in the third feature vector according to the weight value in the weight vector to obtain a weighted feature vector; and splicing the first characteristic vector, the second characteristic vector and the weighted characteristic vector to obtain the average risk influence characteristic. According to the embodiment of the application, the updated weight vector is maintained in real time, the unemployment rates of different industries can be weighted according to the development trends of different industries, the development conditions of the unemployment rates of different industries in the future can be described more accurately, more accurate unemployment rate information is obtained, the target user carries out loan risk prediction under the average risk level according to the more accurate unemployment rate information, and the accuracy of loan risk prediction under the average risk level can be improved.

Wherein the loan risk level comprises a working industry risk level, the user data comprises industry related data, the loan risk prediction model comprises a working industry loan risk prediction model corresponding to the working industry risk level,

the step of performing loan risk prediction at different loan risk levels respectively by using the user data and the loan risk prediction models corresponding to the different loan risk levels comprises:

step B10, constructing a practitioner industry risk level influence characteristic corresponding to the practitioner industry risk level according to the industry related data, wherein the industry related data at least comprises one of industry average income, industry average employment duration and industry number;

and step B20, inputting the professional industry risk level influence characteristics into the professional industry loan risk prediction model, and performing loan risk prediction on the target user under the professional industry risk level.

In this embodiment, it should be noted that the industry-related risk level influence feature is a result that is composed of data in industry-related data and can influence the industry-related risk level, for example, the influence feature of the industry-related risk level at least includes one of industry average income data, industry average employment duration information, and information on number of enterprises in the industry, where the industry average income data may be time-series data for describing a development situation of the industry in recent years, such as an average income level situation of the industry in recent 3 years, or an average income level situation of the industry in recent 5 years, and the industry average income data may describe income index parameters of industry average income level variation, such as an average value, a variance, a standard deviation, or the like; the average employment duration of the industry can be an employment duration index and is used for describing employment stability conditions of the industry, for example, when the employment duration is not less than 5 years, employment can be judged as stable, and when the employment duration is less than 5 years, employment can be judged as unstable; the industry and enterprise quantity data may be time series data for describing changes of the industry in recent years, for example, the changes of the enterprise quantity in the last 5 years, or the changes of the enterprise quantity in the last 10 years, and the industry and enterprise quantity data may describe an enterprise quantity index of the changes of the enterprise quantity, may be a total value of the enterprise quantity, or may be a change value of the enterprise quantity.

As an example, steps B10 to B20 include: acquiring average income data of a professional industry, average employment duration information of the industry and quantity data of industry enterprises; weighting the industry average income data according to preset industry weight information to obtain industry average income data weighting information; according to the weighted average income information of the working industry, the average employment duration information of the industry and the quantity information of the industry enterprises, constructing influence characteristics of the working industry risk level corresponding to the working industry risk level; and performing loan risk prediction on the target user under the professional industry risk level by inputting the professional industry risk level influence characteristics into the professional industry loan risk prediction model.

As an example, the industry-wide average revenue data may be a fourth feature vector comprising a plurality of revenue index parameters, the industry employment duration may be a fifth feature vector comprising a plurality of employment duration index parameters, the industry-wide business quantity information may be a sixth feature vector comprising a business quantity index, the preset industry weight information may be a weight vector composed of weight values corresponding to different indicators of the same industry, where the weight value represents a change of an average income of a target user in a industry, and the higher the industry average income is, the lower the weight value is, for example, if the average income of the current industry of the internet of things is in an increasing trend, the corresponding weight value may be set to 0.5, if the average income of the current industry of the internet of things is in a decreasing trend, the corresponding weight value may be set to 1, if the average income of the current industry of the internet of things is in an increasing trend, the corresponding weight value may be set to 0.5, but when the average income of the industry of the handicraft industry is in a decreasing trend, the corresponding weight value may be set to 1, so step B10 may further include:

and splicing the fourth feature vector, the fifth feature vector and the sixth feature vector to obtain the risk level influence features of the industry of operation. According to the embodiment of the application, the updated weight vector is maintained in real time, the industry income indexes of the working industry can be weighted according to the development trend of the industry, the income condition of the target user can be described more accurately, the loan risk prediction is carried out on the target user under the working risk level, and the accuracy of the loan risk prediction under the working risk level can be improved.

Wherein the loan risk hierarchy comprises an individual risk hierarchy, and the user data comprises at least one of personal credit, personal income, and age,

step C10, constructing individual risk influence characteristics corresponding to the individual risk levels according to the individual credit, the individual income and the age;

and step C20, inputting the individual risk influence characteristics into an individual loan risk prediction model, and performing loan risk prediction on the target user under the individual risk level.

In this embodiment, it should be noted that the individual risk influencing characteristic timely influences the individual risk, and the individual risk influencing characteristic influences the result of the average risk hierarchy, for example, the individual risk influencing characteristic at least includes one of personal credit information, personal income data, and age, where the personal credit information may be a credit indicator for describing an integrity level for a target, for example, when there is no bad record in the personal credit information, the integrity level may be good, when there is a small number of bad records in the personal credit information, the integrity level may be general, and when there is a large number of bad records in the personal credit information, the integrity level may be poor; the personal income data can be a personal income index and can describe the repayment capacity of the target user, for example, when the income level of the target user is higher than the average income level of the region, the target user can be considered to have good repayment capacity, and if the income level of the target user is not higher than the average income level of the region, the target user can be considered to have overdue risk.

As an example, the steps C10 to C20 include: acquiring personal credit investigation information, personal income information and the age of a target user; weighting the personal credit information according to preset credit weight information to obtain a weighted credit index; constructing personal risk influence characteristics corresponding to the personal risk according to the weighted credit index, the income index and the age; and performing loan risk prediction on the target user under the individual risk level by inputting the individual risk influence characteristics into an individual loan risk prediction model.

As an example, the credit investigation information may be a first credit feature vector formed by a plurality of credit records of the user, the personal income information may be a first income feature vector formed by data related to the industry of the target user, the preset credit weight information may be a weight vector formed by weight values corresponding to the plurality of credit records of the target user, the weight values are used for representing the credit condition of the target user, the weight values are lower when the credit condition of the target user is better, and the weight values are higher when the credit condition of the target user is worse, for example, when the target user has a plurality of bad credit records, the weight may be set to 2, and when the target user does not have a bad credit record, the weight value may be set to 0.5; and splicing the first credit characteristic vector, the first income characteristic vector and the age of the target user to obtain individual risk influence characteristics.

According to the credit investigation method and the credit investigation system, the credit investigation weight vector is continuously maintained and updated according to the target user, the credit index of the target user can be weighted according to the credit investigation history and income condition of the target user, the repayment condition of the target user in the future can be more accurately described, and the accuracy of loan risk prediction under the individual risk level can be improved.

The loan risk prediction result comprises a loan risk assessment value, and the step of detecting whether the target user has a loan risk or not according to each loan risk prediction result comprises the following steps:

step S31, carrying out weighted summation on each loan risk assessment value to obtain a total loan risk assessment value;

step S32, if the total loan risk assessment value is smaller than a preset risk threshold, judging that the target user does not have loan risk;

and S33, if the total loan risk assessment value is not less than a preset risk threshold, determining that the target user has a loan risk.

In this embodiment, it should be noted that the loan risk assessment value is a result of loan risk prediction, the obtained corresponding loan risk assessment value can explain the influence of the corresponding risk, and the loan risk assessment value may be a first loan risk probability, a second loan risk probability and a third loan risk probability or a first loan risk level, a second loan risk level and a third loan risk level; the total loan risk assessment value can be the weighted sum of the first loan risk probability, the second loan risk probability and the third loan risk probability or the weighted sum of the first loan risk level, the second loan risk level and the third loan risk level, and the risk of the borrower can be judged.

As an example, the steps S31 to S33 include: according to the first loan risk probability, the second loan risk probability and the third loan risk probability; weighting the first loan risk probability, the second loan risk probability and the third loan risk probability according to preset risk level weight information to obtain weighted loan risk probability; the preset risk level information may be a weight vector composed of weight values corresponding to different risk levels, where the weight vector is used to represent loan risks of different risk levels, and when the loan risk probability is higher, the weight is higher, and if the loan risk probability is lower, the weight is lower, for example, when the first loan risk probability is greater than the second loan risk probability and is greater than the third loan risk probability, and the second loan risk probability is greater than the third loan risk probability, the weight value of the first loan risk probability may be set to 2, the weight value of the second loan risk probability may be set to 1, the weight value of the first loan risk probability may be set to 0.5, and when the first loan risk probability is less than the second loan risk probability and is less than the second loan risk probability, the weight value of the first loan risk probability may be set to 0.5, the weight value of the second loan probability may be set to 1, and the weight value of the first loan risk probability may be set to 2; comparing the loan risk total evaluation value with the preset risk threshold value, and comparing the loan risk total probability with the risk threshold value to obtain a loan risk result of the target user; if the total loan risk probability is higher than the preset risk threshold, it can be determined that the user has a loan risk, and if the total loan risk evaluation valuation is lower than the preset risk threshold, it can be determined that the user does not have a loan risk.

As an example, based on the first loan risk level, the second loan risk level, and the third loan risk level; weighting the first loan risk probability, the second loan risk level and the third loan risk level according to preset risk level weight information to obtain a weighted loan risk level; the preset risk level information may be a weight vector composed of weight values corresponding to different risk levels, where the weight vector is used to represent loan risks of different risk levels, and when a loan risk level is higher, the weight is higher, and if the loan risk level is lower, the weight is lower, for example, when a first loan risk level is higher than a second loan risk level and higher than a third loan risk level, and the second loan risk level is higher than the third loan risk level, the weight value of the first loan risk level may be set to 2, the weight value of the second loan risk level may be set to 1, the weight value of the first loan risk level may be set to 0.5, when the first loan risk level is lower than the second loan risk level and lower than the second loan risk level, the weight value of the first loan risk level may be set to 0.5, the weight value of the second loan risk level may be set to 1, and the weight value of the first loan risk level may be set to 2; comparing the loan risk total evaluation value with the preset risk threshold value, wherein the preset risk threshold value can be used for calculating the risk threshold value according to the known income expectation of the risk of the borrower, and comparing the loan risk total grade with the risk threshold value to obtain a loan risk result of the target user; if the total loan risk level is higher than the preset risk threshold, it can be determined that the user has a loan risk, and if the total loan risk evaluation valuation is lower than the preset risk threshold, it can be determined that the user does not have a loan risk.

Wherein, before the step of determining that the target user does not have a loan risk if the total borrower risk assessment value is smaller than a preset risk threshold, the loan risk detection method comprises:

step D10, obtaining the loan risk information and income expectation of at least one loaned user;

and D20, setting the preset risk threshold according to the risk information of each borrower and the income expectation.

In this embodiment, it should be noted that the borrower risk information of the borrowed user refers to a loan risk total evaluation valuation of the borrowed person, which is used for describing the loan risk of the borrowed person, for example, at least one borrower risk information may include loan risk probabilities or loan risk levels of a plurality of borrowed persons; the preset risk threshold may be a risk threshold parameter for determining whether a borrower has a loan risk, and the risk threshold parameter may be an average value, a variance, or a standard deviation, where the risk threshold parameter may be calculated by a loan risk probability of the borrowed or a loan risk level and a income expectation of the borrowed, for example, 100 ten thousand dollars of interest can be obtained when the loan amount of the borrowed is 300 ten thousand dollars according to a due loan risk probability of the borrowed and a loan level of a target user, and the loan amount, the interest and the income expectation are substituted into formula 1 according to a loan expectation of the borrowed being set to 0, where formula 1 is income expectation = interest (1-risk threshold) -loan amount risk threshold, and the risk threshold is 0.25 according to formula 1, so as to obtain the risk threshold parameter.

As an example, steps D10 to D20 include: obtaining loan risk probability or loan risk level and income expectation of one or more loaned users; according to the preset risk threshold, whether the target user has a loan risk is judged, for example, when the loan probability or the loan grade of a borrower is higher than the preset risk threshold parameter, the borrower has the loan risk, and when the loan probability or the loan grade of the borrower is not higher than the preset risk threshold parameter, the borrower does not have the loan risk.

After the step of determining that the target user has a loan risk if the total loan risk assessment value is not less than a preset risk threshold, the method includes:

step D11, obtaining a weighted summation weight corresponding to the loan risk assessment value;

and D21, determining factors influencing the total evaluation value of the loan risk according to the weighted sum weight and the evaluation values of different loan risk levels.

In this embodiment, it should be noted that the weighted summation weight corresponding to the loan risk assessment value refers to weights corresponding to different risk levels, for example, the weighted summation weight may be a first loan risk probability weight, a second loan risk probability weight, and a third loan risk probability weight, or the first loan risk level weight, the second loan risk level weight, and the third loan risk level weight, and may be used to describe the magnitude of the loan risk assessment value influenced by different levels of risk; the different loan risk level assessment values refer to loan risk results of different levels and can describe the risk levels of the different risk levels, and the different loan risk level assessment values may include a first loan risk probability, a second loan risk probability and a third loan risk probability or a first loan risk level, a second loan risk level and a third loan risk level.

As an example, steps D11 to D21 include: and obtaining the loan risk assessment value, obtaining the first loan risk probability weight, the second loan risk probability weight and the third loan risk probability weight, and obtaining factors influencing the total loan risk assessment value according to the first loan risk probability weight, the second loan risk probability weight and the third loan risk probability weight, wherein if the first loan risk probability weight is 2, the second loan risk probability weight is 1 and the third loan risk probability weight is 0.5, the first loan risk probability value can be determined as the first loan risk probability to the maximum extent, and then the feature vectors of the factors influencing the first risk probability are further determined according to the maximum feature vector weight dynamically set in the first loan risk probability.

As an example, the loan risk assessment value is obtained, the first loan risk level weight, the second loan risk level weight, and the third loan risk level weight are obtained, and the factors affecting the total loan risk assessment value are obtained according to the first loan risk level weight, the second loan risk level weight, and the third loan risk level weight, for example, if the first loan risk level weight is 2, the second loan risk level weight is 1, and the third loan risk level weight is 0.5, the factor affecting the total loan risk assessment value may be determined as the first loan risk level at the maximum, and then the feature vector of the factors affecting the first risk level is further determined according to the maximum feature vector weight dynamically set in the first loan risk level. According to the method and the device, the weights of different loan risk levels are dynamically set according to risk results of different levels, influence factors on loan risk total evaluation valuations of different target users can be more accurately described, and the interpretability of the loan risk total evaluation valuations influencing the target users can be improved.

Compared with the technical means of loan risk prediction based on a traditional machine learning model or loan risk prediction based on a deep learning model adopted in the prior art, the loan risk detection method based on the deep learning model firstly obtains user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of an average risk level, a practice industry risk level deviation level and an individual risk deviation level, so that loan risk prediction is respectively carried out in different loan risk levels by utilizing the user data and the loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results of different loan risk levels, and the loan risk of the target user is respectively evaluated from the average risk level, the practice industry risk level deviation and the individual risk deviation, and whether the target user has a loan risk is comprehensively detected according to the loan risk prediction results. The loan risk prediction model is a deep learning model, so that the accuracy of loan risk prediction results of three levels, namely average risk, industry-of-practice risk level deviation and individual risk deviation, can be ensured, the accuracy of detecting whether the target user has loan risk can be ensured, and when the loan risk exists, the influence degree of the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation, on the final detection of whether the target user has loan risk can be explained according to the loan risk prediction results of the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation, so that the interpretability of the loan risk detection on the three levels, namely the average risk, the industry-of-practice risk level deviation and the individual risk deviation is ensured.

Example two

Referring to fig. 2, an embodiment of the present application further provides a loan risk detection apparatus, where the loan risk detection apparatus includes:

the system comprises an acquisition module 10, a processing module and a processing module, wherein the acquisition module is used for acquiring user data of a target user in different loan risk levels, and the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level;

the risk prediction module 20 is used for performing loan risk prediction at different loan risk levels by using the user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results at the different loan risk levels;

and the risk detection module 30 is used for detecting whether the target user has loan risks or not according to the loan risk prediction results.

Optionally, the loan risk hierarchy includes an average risk hierarchy, the user data includes macro economic indicator data, the loan risk prediction model includes an average loan risk prediction model corresponding to the average risk, and the risk prediction module 20 is further configured to:

Optionally, the loan risk level includes a practice industry risk level, the user data includes industry-related data, the loan risk prediction model includes a practice industry loan risk prediction model corresponding to the practice industry risk level, and the risk prediction module 20 is further configured to:

according to the industry related data, constructing a industry risk level influence characteristic corresponding to the industry risk level, wherein the industry related data at least comprises one of industry average income, industry average employment duration and the number of enterprises in the industry;

Optionally, the loan risk hierarchy comprises an individual risk hierarchy, the user data comprises at least one of personal credit, personal income, and age, and the risk prediction module 20 is further configured to:

Optionally, the loan risk prediction result includes a loan risk assessment value, and the risk detection module 30 is further configured to:

Optionally, the loan risk detection apparatus is further configured to:

By adopting the loan risk detection method in the first embodiment, the loan risk detection device provided by the application solves the technical problem that the accuracy and the interpretability of loan risk prediction cannot be considered at the same time. Compared with the prior art, the beneficial effects of the loan risk provided by the embodiment of the present application are the same as the beneficial effects of the loan risk detection method provided by the above embodiment, and other technical features of the loan risk detection device are the same as those disclosed in the above embodiment method, which are not described herein again.

EXAMPLE III

An embodiment of the present application provides an electronic device, where the electronic device may be a loan risk, and the electronic device includes: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the loan risk detection method of the first embodiment.

Referring now to FIG. 3, shown is a schematic diagram of an electronic device suitable for use in implementing embodiments of the present disclosure. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., car navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 3, the electronic device may include a processing means (e.g., a central processing unit, a graphic processor, etc.) that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) or a program loaded from a storage means into a Random Access Memory (RAM). In the RAM, various programs and data necessary for the operation of the electronic apparatus are also stored. The processing device, the ROM, and the RAM are connected to each other by a bus. An input/output (I/O) interface is also connected to the bus.

Generally, the following systems may be connected to the I/O interface: input devices including, for example, touch screens, touch pads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, and the like; output devices including, for example, liquid Crystal Displays (LCDs), speakers, vibrators, and the like; storage devices including, for example, magnetic tape, hard disk, etc.; and a communication device. The communication means may allow the electronic device to communicate wirelessly or by wire with other devices to exchange data. While the figures illustrate an electronic device with various systems, it is to be understood that not all illustrated systems are required to be implemented or provided. More or fewer systems may alternatively be implemented or provided.

In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means, or installed from a storage means, or installed from a ROM. The computer program, when executed by a processing device, performs the above-described functions defined in the methods of the embodiments of the present disclosure.

The electronic device provided by the application adopts the loan risk detection method in the first embodiment to solve the technical problem that the accuracy and the interpretability of the loan risk prediction cannot be considered at the same time. Compared with the prior art, the beneficial effects of the loan risk provided by the embodiment of the present application are the same as the beneficial effects of the loan risk detection method provided by the above embodiment, and other technical features of the loan risk detection device are the same as those disclosed in the above embodiment method, which are not described herein again.

It should be understood that portions of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof. In the foregoing description of embodiments, the particular features, structures, materials, or characteristics may be combined in any suitable manner in any one or more embodiments or examples.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Example four

The present embodiment provides a computer-readable storage medium having computer-readable program instructions stored thereon for performing the method for loan risk detection in the first embodiment.

The computer readable storage medium provided by the embodiments of the present application may be, for example, a usb disk, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, or device, or a combination of any of the above. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present embodiment, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, or device. Program code embodied on a computer readable storage medium may be transmitted using any appropriate medium, including but not limited to: models, mean risk hierarchies, professional and individual risk hierarchies, and the like, or any suitable combination of the foregoing.

The computer-readable storage medium may be embodied in an electronic device; or may be present alone without being incorporated into the electronic device.

The computer-readable storage medium carries one or more programs which, when executed by an electronic device, cause the electronic device to: acquiring user data of a target user in different loan risk levels, wherein the different loan risk levels at least comprise one of an average risk level, a professional risk level and an individual risk level; performing loan risk prediction at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results at the different loan risk levels; and detecting whether the target user has loan risks or not according to the loan risk prediction results.

Computer program code for carrying out operations for aspects of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules described in the embodiments of the present disclosure may be implemented by software or hardware. Wherein the names of the modules do not in some cases constitute a limitation of the unit itself.

The computer readable storage medium provided by the application stores computer readable program instructions for executing the loan risk detection method, and solves the technical problem that the accuracy and interpretability of loan risk prediction cannot be considered at the same time. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the embodiment of the present application are the same as the beneficial effects of the loan risk detection method provided by the first embodiment, and are not described herein again.

EXAMPLE five

The present application also provides a computer program product comprising a computer program which, when executed by a processor, performs the steps of the loan risk prediction method as described above.

The computer program product provided by the application solves the technical problem that the accuracy and the interpretability of loan risk prediction cannot be considered at the same time. Compared with the prior art, the beneficial effects of the computer program product provided by the embodiment of the present application are the same as the beneficial effects of the loan risk detection method provided by the first embodiment of the present application, and are not described herein again.

The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all equivalent structures or equivalent processes, which are directly or indirectly applied to other related technical fields, and which are not limited by the present application, are also included in the scope of the present application.

Claims

1. A loan risk detection method is characterized by comprising the following steps:

2. The loan risk detection method of claim 1, wherein the loan risk hierarchy includes an average risk hierarchy, the user data includes macro economic indicator data, the loan risk prediction model includes an average loan risk prediction model corresponding to the average risk,

the step of performing loan risk prediction at different loan risk levels respectively by using the user data and loan risk prediction models corresponding to the different loan risk levels comprises:

3. The loan risk detection method of claim 1, wherein the loan risk hierarchy includes a working industry risk hierarchy, the user data includes industry-related data, the loan risk prediction model includes a working industry loan risk prediction model corresponding to the working industry risk hierarchy,

4. The loan risk detection method of claim 1, wherein the loan risk tiers include individual risk tiers, the user data includes at least one of personal credit, personal income, and age,

5. The loan risk detection method according to claim 1, wherein the loan risk prediction result includes a loan risk assessment value, and the step of detecting whether the target user has a loan risk based on each loan risk prediction result includes:

6. The loan risk detection method of claim 5, wherein before the step of determining that the target user does not have a loan risk if the total borrower risk assessment value is less than a preset risk threshold, the loan risk detection method further comprises:

acquiring the borrower risk information and income expectation of at least one borrowed user;

7. The loan risk detection method of claim 5, wherein after the step of determining that the target user is at a loan risk if the total borrower risk assessment value is not less than a preset risk threshold, the method further comprises:

and determining factors influencing the total evaluation value of the loan risk according to the weighted sum weight and the evaluation values of different loan risk levels.

8. A loan risk detection apparatus, comprising:

the risk prediction module is used for performing loan risk prediction on different loan risk levels by using user data and loan risk prediction models corresponding to the different loan risk levels to obtain loan risk prediction results of the different loan risk levels;

9. An electronic device, characterized in that the electronic device comprises:

at least one processor;

and a memory communicatively coupled to the at least one processor;

wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the loan risk detection method of any of claims 1 to 7.

10. A computer-readable storage medium, having stored thereon a program for implementing a loan risk detection method, the program being executable by a processor to implement the steps of the loan risk detection method according to any one of claims 1 to 7.