CN114418729A

CN114418729A - White household default risk assessment method and device, electronic equipment and storage medium

Info

Publication number: CN114418729A
Application number: CN202111491284.5A
Authority: CN
Inventors: 席颐画; 宋万鹏
Original assignee: Tongdun Network Technology Co ltd
Current assignee: Tongdun Network Technology Co ltd
Priority date: 2021-12-08
Filing date: 2021-12-08
Publication date: 2022-04-29

Abstract

The application relates to a white household default risk assessment method, a white household default risk assessment device, electronic equipment and a storage medium, and belongs to the technical field of risk prediction, wherein the method comprises the following steps: acquiring white user attribute characteristics, and matching non-white user sets with the same or similar attribute characteristics; predicting default risk scores of samples in a non-white household set through a pre-constructed default risk prediction model; and calculating the default risk score of the white house by a time weighted average method based on the default risk score. According to the embodiment of the application, the default risk score of the white household can be calculated based on the prediction score of the non-white household set without depending on expert suggestions under the condition that white household data are missing, the risk assessment of the white household is more objective and accurate, and the method and the device have wide applicability.

Description

White household default risk assessment method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of risk prediction technologies, and in particular, to a white household default risk assessment method, apparatus, electronic device, and storage medium.

Background

The "white user" refers to a user who has no data other than the attribute features among all data that can be acquired by the financial institution. Particularly, if the financial institution is a banking institution, the bank white client refers to a client who has not transacted any loan or credit card of the financial institution and has not entered personal information by the credit investigation center of the people's bank.

Due to the data loss of the white user, most of domestic financial institutions such as banks and the like cannot evaluate default risks of the white user at present, and the white user risk evaluation can be a great difficulty for various financial institutions to perfect a user risk evaluation system. The existing scheme is based on attribute characteristics and expert suggestions, a set of strategies is independently formulated for white users, the white users are simply divided into credit grades by combining the strategies, the method is rough and subjective, wide applicability is lacked, and the accuracy of an evaluation result is low.

Disclosure of Invention

The embodiment of the application provides a white household default risk assessment method, a white household default risk assessment device, electronic equipment and a storage medium, and aims to at least solve the problems that credit rating classification is rough and subjective, wide applicability is lacked, and the accuracy of assessment results is low in the related technology.

In a first aspect, an embodiment of the present application provides a white household default risk assessment method, including: acquiring white user attribute characteristics, and matching non-white user sets with the same or similar attribute characteristics; predicting default risk scores for the samples in the non-white household set through a pre-constructed default risk prediction model; and calculating the default risk score of the white household by a time weighted average method based on the default risk score.

In some embodiments, the matching out the non-white user sets with the same or similar attribute characteristics includes: acquiring non-white house samples with the same attribute characteristics as N white houses within a preset time before the white house loan application date to obtain a non-white house set, wherein N is a positive integer greater than 1; if the number of samples in the non-white user set is smaller than a set threshold value, removing M attribute features, wherein M is larger than or equal to 1 and is smaller than N; matching non-white user sets with the same attribute characteristics according to the remaining attribute characteristics; repeating the culling and matching steps until the number of samples in the non-white user set is greater than or equal to a set threshold.

In some embodiments, the culling M attribute features comprises: and eliminating M attribute features with minimum variable importance.

In some of these embodiments, the default risk prediction model includes a first layer model and a second layer model, the first layer model includes a plurality of sub-models, the sub-models are used for predicting results according to data of corresponding dimensions, the second layer model is a fusion model, and the predicting default risk scoring of the samples in the non-white household set through the pre-constructed default risk prediction model includes: respectively inputting the data of each dimension into each submodel to obtain a plurality of predicted results; and inputting the results of the multiple predictions into the fusion model for fusion to obtain a prediction default risk score.

In some embodiments, the step of determining M attribute features with the least importance of the variables includes: and acquiring the importance of the model entering variables according to each submodel, and sequencing according to the sizes to obtain M attribute characteristics with the minimum importance of the variables.

In some of these embodiments, the second layer model is a logistic regression model.

In some embodiments, the calculating the default risk score of the white household by a time weighted average method based on the default risk score comprises:

wherein f is_t(x) The default risk score of the white household x on the loan application date T is represented, T represents the maximum month that the loan application date of the sample in the non-white household set is far away from the loan application date T of the white household,

and (4) representing the mean value of matched non-white household sample default risk scores in the i month before the loan application date t.

In a second aspect, an embodiment of the present application provides a white household default risk assessment apparatus, including: the device comprises a matching module, a prediction module and a calculation module, wherein the matching module is used for acquiring the attribute characteristics of the white users and matching non-white user sets with the same or similar attribute characteristics; the prediction module is used for predicting default risk scores of the samples in the non-white household set through a pre-constructed default risk prediction model; and the calculation module is used for calculating the default risk score of the white household by a time weighted average method based on the default risk score.

In a third aspect, an embodiment of the present application provides an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor is configured to execute the computer program to perform any one of the methods described above.

In a fourth aspect, an embodiment of the present application provides a storage medium, in which a computer program is stored, where the computer program is configured to execute any one of the methods described above when the computer program runs.

Compared with the related art, the white household default risk assessment method provided by the embodiment of the application can be used for calculating the default risk score of the white household based on the forecast default risk score of the non-white household set without depending on expert suggestions under the condition that white household data are missing, is more objective and accurate in risk assessment of the white household, and has wide applicability. Firstly, when non-white users with the same or similar attribute characteristics are matched with the white users, if the number of samples is not enough, the least attribute characteristics are removed every time for matching, and the accuracy of the matching result can be guaranteed to the greatest extent, so that the accuracy of the final prediction result is improved. Secondly, the default risk prediction model of the embodiment of the application adopts a double-layer model, and has the advantages that the combination of different data dimensions can be flexibly carried out, if a certain data dimension changes, other submodels are not affected, and only the submodel and the fusion model with changed data need to be adjusted. In addition, the prediction results of the sub-models are fused through the fusion model, so that the final prediction result is more accurate, wherein the fusion model adopts a logistic regression model, and the number of the sub-models is preferably within 20 in the actual service, so that the probability of overfitting of the default risk prediction model can be effectively reduced by using the logistic regression model under the condition of less model entering variables. In addition, the embodiment of the application also adopts a time weighted average method, the influence of time factors on the result is considered, and the accuracy of the evaluation result is greatly improved on the whole.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

FIG. 1 is a flowchart of a white house default risk assessment method according to an embodiment of the present application;

FIG. 2 is a schematic structural diagram of a default risk prediction model according to an embodiment of the present application;

FIG. 3 is a schematic structural diagram of a white household default risk assessment device according to an embodiment of the present application;

fig. 4 is an internal structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.

Reference in the specification 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 specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.

Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Reference herein to "a plurality" means greater than or equal to two. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.

FIG. 1 is a flowchart of a white user default risk assessment method according to an embodiment of the present application, as shown in FIG. 1, the method includes the following steps;

s101: acquiring white user attribute characteristics, and matching non-white user sets with the same or similar attribute characteristics;

s102: predicting default risk scores of samples in a non-white household set through a pre-constructed default risk prediction model;

s103: and calculating the default risk score of the white house by a time weighted average method based on the default risk score.

According to the content, the default risk score of the white user can be calculated based on the forecast default risk score of the non-white user set without depending on expert suggestions under the condition that white user data are missing.

For the purpose of more clearly explaining the present application, specific examples are set forth below in detail.

Firstly, aiming at a default risk prediction model, a construction mode is provided as follows:

(1) acquiring non-white-user multidimensional data as training data, for example, dimension 1 data including attribute features and borrowing data, wherein the attribute features include age information, gender information, province information and the like; dimension 2 data including device data, such as information used by a user on a mobile terminal, a PC terminal, and the like; dimension 3 data, including operator data; dimension n data, including e-commerce logistics data.

(2) The structure of the default risk prediction model is designed to include a first layer model and a second layer model, fig. 2 is a schematic structural diagram of the default risk prediction model according to the embodiment of the application, as shown in fig. 2, the first layer model is composed of a plurality of submodels, in the training stage, each submodel is trained based on data of corresponding dimension, for example, the submodel 1 is trained based on data of dimension 1, and the submodel 2 is trained based on data of dimension 2, so that a trained submodel is finally obtained in each dimension; in the application phase, each submodel predicts the result from the data of its respective dimension. In addition, each sub-model may employ a gradient lifting tree, such as XGBoost, LightGBM; the second layer model is a fusion model, such as a logistic regression model.

(3) When the default risk prediction model is trained, firstly, the submodels of the first layer model are trained, the submodels are independent from each other, and the submodels are respectively trained, for example, the submodel 1 is trained firstly, and the submodel 1 is subjected to parameter adjustment; then, training the sub-model 2, performing parameter adjustment on the sub-model 2, and sequentially training and adjusting parameters of the other sub-models; alternatively, the training sequence may be randomized. And then, after all the sub models are trained completely, constructing a second layer of model, namely a logistic regression model.

As an example, the training process of the submodel includes: a. determining modeling samples, extracting 60% of the modeling samples for training the submodels, and selecting corresponding data dimension variables as parameter-entering variables (such as constructing a submodel 1, wherein the parameter-entering variables comprise attribute characteristics and borrowing data; such as constructing a submodel 2, and the parameter-entering variables comprise equipment data); b. primarily screening the variables, wherein the screening standard can be formulated according to the actual situation; c. carrying out model training and parameter adjustment; d. and obtaining the trained sub-model.

The training process of the second layer model comprises the following steps: a. using the remaining 40% of samples as modeling samples of the logistic regression model; b. predicting the remaining 40% of samples based on the n sub-models of the first layer model, and taking the predicted result as an input variable of the logistic regression model; c. carrying out model training and parameter adjustment; d. and obtaining the trained logistic regression model. It should be noted that, a person skilled in the art knows the training modes of the gradient lifting tree and the logistic regression model, so that the person skilled in the art can construct the first layer model and the second layer model when knowing the content of the present application, and therefore, the training process of the model is not described in detail here.

Based on the above contents, n pieces of dimensional data are respectively input into n pieces of sub-models to obtain n pieces of results, and the n pieces of results are input into a logistic regression model for fusion, so that the finally obtained prediction result is more accurate.

The embodiment of the application adopts the double-layer model, and has the advantages that the combination of different data dimensions can be flexibly carried out, if a certain data dimension is changed, other submodels are not influenced, and only the submodel and the fusion model with changed data need to be adjusted. The fusion model adopts a logistic regression model, and the number of the optimized sub-models in the actual service is within 20, so that the probability of overfitting of the default risk prediction model can be effectively reduced by using the logistic regression model under the condition of less model-entering variables.

When the white user needs to be assessed for the default risk, the method proceeds to step S101: and acquiring the attribute characteristics of the white households, and matching a non-white household set with the same or similar attribute characteristics, wherein the aim is to find out a non-white household sample which is the same as or similar to the white households.

As an example, obtaining a non-white house sample with N attribute characteristics that are the same as the white house within a preset time (for example, 12 months) before the white house loan date to obtain a non-white house set, where N is a positive integer greater than 1; if the number of samples in the non-white user set is smaller than a set threshold value, M attribute features are removed, wherein M is larger than or equal to 1 and is smaller than N; matching a non-white user set according to the remaining attribute features; and repeating the elimination and the matching until the number of the samples in the matched non-white user set is greater than or equal to a set threshold value. It should be noted that the threshold value may be set according to a certain proportion of the monthly average user amount of the financial institution.

Assuming that N is 15 and M is 2, the set threshold is 50, in one case, non-white user samples having the same attribute characteristics as 15 white users are obtained, and the number of samples in the non-white user set is 60 (greater than 50), then the 60 non-white user samples may be considered to be the same as the white users; in another case, non-white user samples with the same attribute characteristics as 15 white users are obtained, and the number of the samples in the non-white user set is 20 (less than 50), which means that the matching condition is harsh, therefore, the number of the matched samples is small, and therefore, 2 attribute features can be eliminated, at the moment, non-white user samples which are the same as 13 attribute features of the white user are obtained, and due to the fact that the matching condition is relaxed, therefore, the number of samples in the non-white user set is greater than 20, for example, 40, but the number of samples is still less than 50, for example, then 2 attribute features need to be eliminated, at this time, non-white user samples identical to 11 attribute features of the white user are obtained, because the matching condition is relaxed, the number of samples in the obtained non-white user set is greater than 40, and therefore, repeating the steps of removing and matching will eventually obtain greater than or equal to 50 non-white user samples. For example, when 9 attribute features remain, 51 non-white user samples are matched to meet the requirement that the number of samples is greater than the set threshold, and the 51 non-white user samples can be considered to be similar to white users.

Further, M attribute features with the minimum variable importance are removed each time. It should be noted that, when constructing the default risk prediction model, a sub-model is constructed based on different data dimensions, and the importance of the modeling variables (i.e., attribute features) can be obtained through the constructed sub-model, so that the importance of the variables can be sorted according to size, and the M attribute features with the minimum importance of the variables are selected. For example, the variable importance is arranged from large to small, and the last M attribute features are selected for elimination. If M is equal to 1, the last attribute feature is rejected, if the variable importance of M attribute features is equal and is the minimum value, the M attribute features may be rejected, or any one of the M attribute features may be rejected. Therefore, the accuracy of the matching result can be guaranteed to the maximum extent as the least important attribute features are removed each time.

In the embodiment of the present application, it is considered that the attribute characteristics of the white household may also change with time, for example, the income of the white household changes or the consumption concept changes, and the change can be reflected by the transaction data of different periods, so the embodiment of the present application mainly focuses on the non-white household which is the same as or similar to the white household within a preset time (for example, 12 months) before the white household loan application date, and can more accurately reflect the latest real condition of the white household.

After matching out the non-white user set, the process proceeds to step S102: and predicting default risk scores of the samples in the non-white household set through a pre-constructed default risk prediction model. Assuming that 51 non-white household samples exist in the non-white household set, multi-dimensional data (including attribute characteristics, borrowing data, equipment data, operator data, e-commerce logistics data and the like) of the 51 non-white household samples are input into the default risk prediction model, and 51 predicted default risk scores are obtained.

Subsequently, the flow proceeds to step S103: and calculating the default risk score of the white house by a time weighted average method based on the default risk score. For example, calculating the default risk score f of the white house x on the loan application date t_t(x) Comprises the following steps:

wherein, T tableThe maximum month of the loan application date of the sample in the non-white household set from the loan application date t of the white household is generally taken in actual business

For example, the white household applies for a loan at 11/18 th of 2021, 4 non-white households in the non-white household set apply for loans within 12 th of 11/18 th of 2021 (i.e., 11/1 th of 2020-11/30 th of 2020), the predicted default risk scores of the 4 non-white households are 41, 45, 38 and 42, respectively, and the average value of the non-white household sample default risk scores matched within 12 th of months before the loan application date of the white household is (41+45+38+42)/4 ═ 41.5;

for another example, 6 non-white households in the non-white household set apply for loans within 11 th month before 11 th month and 18 th month in 2021 (i.e., 12 th month and 1 st in 2020 to 12 th month and 31 st in 2020), the predicted default risk scores of the 6 non-white households are 43, 37, 44, 41, 38 and 37 respectively, and then the average value of the matched non-white household sample default risk scores within 11 th month before the loan application date of the white households is (43+37+44+41+38+37)/6 ═ 40;

in the above manner, each can be calculated

And the value of T is known, so f can be finally calculated_t(x) The value of (c).

In consideration of the fact that the multidimensional data of the non-white client may change along with the change of time, the embodiment of the present application mainly focuses on the non-white client who applies for the loan in the near term (for example, the last 12 months) of the loan application date of the white client; meanwhile, the embodiment of the application also considers the influence of the distance between the non-white house loan application date and the white house loan application date (namely, the T, T-1, …, 2,1 month) on the default risk. Therefore, the accuracy of the default risk score of the white user can be greatly improved by the time weighted average method.

It should be noted that the attribute features are mainly used for matching a non-white household set which is the same as or similar to the attribute features of the white household, and specific data of the attribute features depend on feature data of the white household and do not include loan data, credit investigation data and the like; the multidimensional data is mainly used for training each submodel, and the specific data depends on the characteristic data of non-white customers, including loan data, wherein the multidimensional data can contain attribute characteristics.

An embodiment of the present application further provides a white household default risk assessment apparatus, fig. 3 is a schematic structural diagram of the white household default risk assessment apparatus according to the embodiment of the present application, and as shown in fig. 3, the apparatus includes a matching module 1, a prediction module 2, and a calculation module 3. The matching module 1 is used for obtaining the attribute characteristics of the white households and matching non-white household sets with the same or similar attribute characteristics; the prediction module 2 is used for predicting default risk scores of samples in the non-white household set through a default risk prediction model which is constructed in advance; the calculating module 3 is used for calculating default risk scores of the white households by a time weighted average method based on the default risk scores.

The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.

For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and optional implementation manners, and details of this embodiment are not described herein again.

In addition, in combination with the white household default risk assessment method in the foregoing embodiment, the embodiment of the present application may be implemented by providing a storage medium. The storage medium having stored thereon a computer program; the computer program, when executed by a processor, implements any of the above-described embodiments of the method for assessing a white space default risk.

An embodiment of the present application also provides an electronic device, which may be a terminal. The electronic device comprises a processor, a memory, a network interface, a display screen and an input device which are connected through a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic equipment comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the electronic device is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a white house default risk assessment method. The display screen of the electronic equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the electronic equipment, an external keyboard, a touch pad or a mouse and the like.

In one embodiment, fig. 4 is a schematic diagram of an internal structure of an electronic device according to an embodiment of the present application, and as shown in fig. 4, there is provided an electronic device, which may be a server, and its internal structure diagram may be as shown in fig. 4. The electronic device comprises a processor, a network interface, an internal memory and a non-volatile memory connected by an internal bus, wherein the non-volatile memory stores an operating system, a computer program and a database. The processor is used for providing calculation and control capability, the network interface is used for communicating with an external terminal through network connection, the internal memory is used for providing an environment for an operating system and the running of a computer program, the computer program is executed by the processor to realize a white default risk assessment method, and the database is used for storing data.

Those skilled in the art will appreciate that the configuration shown in fig. 4 is a block diagram of only a portion of the configuration associated with the present application, and does not constitute a limitation on the electronic device to which the present application is applied, and a particular electronic device may include more or less components than those shown in the drawings, or combine certain components, or have a different arrangement of components.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

It should be understood by those skilled in the art that various features of the above-described embodiments can be combined in any combination, and for the sake of brevity, all possible combinations of features in the above-described embodiments are not described in detail, but rather, all combinations of features which are not inconsistent with each other should be construed as being within the scope of the present disclosure.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A white household default risk assessment method is characterized by comprising the following steps:

acquiring white user attribute characteristics, and matching non-white user sets with the same or similar attribute characteristics;

predicting default risk scores for the samples in the non-white household set through a pre-constructed default risk prediction model;

and calculating the default risk score of the white household by a time weighted average method based on the default risk score.

2. The method of claim 1, wherein matching out the non-white user sets with the same or similar attribute characteristics comprises:

acquiring non-white house samples with the same attribute characteristics as N white houses within a preset time before the white house loan application date to obtain a non-white house set, wherein N is a positive integer greater than 1;

if the number of samples in the non-white user set is smaller than a set threshold value, removing M attribute features, wherein M is larger than or equal to 1 and is smaller than N;

matching non-white user sets with the same attribute characteristics according to the remaining attribute characteristics;

repeating the culling and matching steps until the number of samples in the non-white user set is greater than or equal to a set threshold.

3. The method of claim 2, wherein the culling M attribute features comprises:

and eliminating M attribute features with minimum variable importance.

4. The method of claim 3, wherein the breach risk prediction model comprises a first layer model and a second layer model, the first layer model comprising a plurality of sub-models for predicting results from data of corresponding dimensions, the second layer model being a fusion model, the predicting breach risk scores for samples in the set of non-whites by a pre-built breach risk prediction model comprising:

respectively inputting the data of each dimension into each submodel to obtain a plurality of predicted results;

and inputting the results of the multiple predictions into the fusion model for fusion to obtain a prediction default risk score.

5. The method according to claim 4, wherein the step of determining the M attribute features with the minimum variable importance comprises:

and acquiring the importance of the model entering variables according to each submodel, and sequencing according to the sizes to obtain M attribute characteristics with the minimum importance of the variables.

6. The method of claim 4, wherein the second layer model is a logistic regression model.

7. The method of claim 1, wherein calculating the default risk score for the white household by a time-weighted average method based on the default risk score comprises:

8. A white home breach risk assessment device, comprising:

the matching module is used for acquiring the attribute characteristics of the white households and matching non-white household sets with the same or similar attribute characteristics;

the prediction module is used for predicting default risk scores of the samples in the non-white household set through a pre-constructed default risk prediction model;

and the calculation module is used for calculating the default risk score of the white household by a time weighted average method based on the default risk score.

9. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 7.

10. A storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any one of claims 1 to 7 when executed.