Scene credit automated modeling method, device, electronic device and storage medium

Abstract

The present disclosure relates to a method, apparatus, electronic device, and storage medium for automated modeling of scenario credit granting, and pertains to the field of Internet of Things technology. The method includes: preprocessing the data of multiple new scenario users to obtain a data set; wherein the data set includes the scenario feature sets of multiple new scenario users; determining one or more target existing scenario credit granting models according to the scenario feature sets of multiple new scenario users and multiple existing scenario credit granting models; obtaining the target scenario feature sets and target credit limits of new scenario users according to the scenario feature sets of new scenario users, the determined one or more target existing scenario credit granting models, and / or the credit limits of multiple existing users in the determined one or more target existing scenario credit granting models; and constructing a new scenario credit granting model according to the target scenario feature sets and target credit limits of multiple new scenario users. Thus, it is possible to automate the modeling of new scenario users, greatly improving the efficiency of establishing credit granting models.

Description

Technical Field

[0001] The present invention relates to the technical field of the Internet of Things, and in particular, to a method, device, electronic device, and storage medium for automatically modeling scenario credit granting. Background Art

[0002] Currently, when financial institutions grant credit to users in different scenarios, different scenario credit granting models are set for different scenarios to grant credit to users through the scenario credit granting models.

[0003] In the related art, when a financial institution needs to introduce new-scenario users, it needs to re-establish a new scenario credit granting model based on the new-scenario users. Re-establishing a new scenario credit granting model requires repeated communication among multiple parties such as the scenario business side, the financial institution business side, and the technical side. From the initial data analysis to model building and finally deployment and application, it takes a large amount of manpower and time. The modeling process is time-consuming and complex, which is not conducive to the financial institution to quickly carry out the credit granting business for new-scenario users. Summary of the Invention

[0004] The present disclosure provides a method, device, electronic device, and storage medium for automatically modeling scenario credit granting.

[0005] The technical solution of the present disclosure is as follows:

[0006] According to the first aspect of the embodiments of the present disclosure, a method for automatically modeling scenario credit granting is provided, including: preprocessing data of multiple new-scenario users to obtain a data set; wherein the data set includes scenario feature sets of multiple new-scenario users; determining one or more target existing scenario credit granting models according to the scenario feature sets of the new-scenario users and multiple existing scenario credit granting models; obtaining the target scenario feature set and target amount of the new-scenario users according to the scenario feature sets of the new-scenario users, the determined one or more target existing scenario credit granting models, and / or the credit granting amounts of multiple stock users in the determined one or more target existing scenario credit granting models; and constructing a new scenario credit granting model according to the target scenario feature sets and target amounts of multiple new-scenario users.

[0007] According to a second aspect of the embodiments of the present disclosure, there is provided a device for automated modeling of scenario credit granting, including: a data set acquisition unit for preprocessing the data of multiple new scenario users to obtain a data set; wherein the data set includes scenario feature sets of multiple new scenario users; a model selection unit for determining one or more target existing scenario credit granting models according to the scenario feature sets of multiple new scenario users and multiple existing scenario credit granting models; a processing unit for obtaining the target scenario feature set and target credit limit of the new scenario users according to the scenario feature sets of the new scenario users, the determined one or more target existing scenario credit granting models, and / or the credit limits of multiple existing users in the determined one or more target existing scenario credit granting models; a model construction unit for constructing a new scenario credit granting model according to the target scenario feature sets and the target credit limits of multiple new scenario users.

[0008] According to a third aspect of the embodiments of the present disclosure, there is provided an electronic device, including: a processor; a memory for storing executable instructions of the processor; wherein the processor is configured to execute the instructions to implement the scenario credit granting automated modeling method proposed in the first aspect of the embodiments of the present disclosure.

[0009] According to a fourth aspect of the embodiments of the present disclosure, there is provided a storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, enabling the electronic device to execute the scenario credit granting automated modeling method proposed in the first aspect of the embodiments of the present disclosure.

[0010] According to a fifth aspect of the embodiments of the present disclosure, there is provided a computer program product, when the computer program product is executed by a processor of an electronic device, enabling the electronic device to execute the scenario credit granting automated modeling method proposed in the first aspect of the embodiments of the present disclosure.

[0011] The technical solutions provided by the embodiments of the present disclosure at least bring the following beneficial effects:

[0012] Preprocess the data of multiple new-scenario users to obtain a data set; wherein, the data set includes the scenario feature sets of multiple new-scenario users; determine one or more target existing-scenario credit models according to the scenario feature sets of multiple new-scenario users and multiple existing-scenario credit models; obtain the target scenario feature set and target credit limit of the new-scenario users according to the scenario feature set of the new-scenario users, the determined one or more target existing-scenario credit models, and / or the credit limits of multiple existing users in the determined one or more target existing-scenario credit models; construct a new-scenario credit model according to the target scenario feature sets and target credit limits of multiple new-scenario users. Thus, it is possible to automatically construct a new-scenario credit model using the data of new-scenario users, saving costs. At the same time, according to the scenario feature set of the new-scenario users and the determined one or more target existing-scenario credit models, the target scenario feature set and target credit limit of the new-scenario users can be obtained to establish a new-scenario credit model, realizing the rapid growth of the business.

[0013] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit the present disclosure. Brief Description of the Drawings

[0014] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure and do not constitute an improper limitation of the present disclosure.

[0015] Figure 1 is a flowchart according to the first embodiment of the present disclosure;

[0016] Figure 2 is a flowchart of the sub-steps of S2 according to the first embodiment of the present disclosure;

[0017] Figure 3 is a flowchart of the sub-steps of S3 according to the first embodiment of the present disclosure;

[0018] Figure 4 is a flowchart of the sub-steps of S32 according to the first embodiment of the present disclosure;

[0019] Figure 5 is a flowchart of the sub-steps of S4 according to the first embodiment of the present disclosure;

[0020] Figure 6 is a structural diagram according to the second embodiment of the present disclosure;

[0021] Figure 7 is a structural diagram of the model selection unit according to the second embodiment of the present disclosure;

[0022] Figure 8 is a structural diagram of the processing unit according to the second embodiment of the present disclosure;

[0023] Figure 9 It is a structural diagram of a first processing subunit according to the second embodiment of the present disclosure;

[0024] Figure 10 It is a structural diagram of a second processing subunit according to the second embodiment of the present disclosure;

[0025] Figure 11 It is a structural diagram of a model construction unit according to the second embodiment of the present disclosure;

[0026] Figure 12 It is a block diagram of an electronic device for implementing the scenario credit automated modeling method according to the embodiment of the present disclosure. Detailed implementation manners

[0027] In order to enable those of ordinary skill in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present disclosure described herein can be implemented in an order different from those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0029] Figure 1 It is a flowchart of a scenario credit automated modeling method shown according to an exemplary embodiment.

[0030] In the embodiments of the present disclosure, the scenario credit automated modeling method is configured in a scenario credit automated modeling device as an example. The scenario credit automated modeling device can be implemented in a software and / or hardware manner, and the scenario credit automated modeling device can be configured in an electronic device so that the electronic device can perform the processing function of tasks.

[0031] Among them, the electronic device can be any device with computing capabilities, such as a PC (Personal Computer), a mobile terminal, a server, etc. The mobile terminal can be, for example, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, a vehicle-mounted device, etc., which are hardware devices with various operating systems, touch screens, and / or display screens.

[0032] In the related technology, credit products are designed with different credit schemes for data from various scenarios. When it is hoped to introduce more users by adding new scenarios, a group of users is usually imported from the new scenario. Based on this group of users and combined with the business characteristics of the new scenario, expert experience analysis is conducted to formulate an established credit formula to be put online for granting credit to users in the scenario.

[0033] However, the problems with this method are: 1. New scenarios need to import enough users to support empirical inductive analysis. When the number of users is small, this method is not enough. 2. The method of summarizing the credit formula entirely through expert experience is subjective and one-sided. 3. When importing one or more scenarios with a small number of users, it is necessary to repeatedly model each scenario, which consumes a lot of manpower and material resources. At the same time, the benefits that scenarios with a small number of users can bring are also limited, and manual modeling is no longer applicable.

[0034] Based on this, a scenario credit automation modeling method is provided in the disclosed embodiment, which can automatically model the imported new scenario users, greatly improving the efficiency of credit model establishment, saving manpower and time costs, and the automatically established credit model can better meet the credit needs of new credit business scenarios.

[0035] The present invention provides a scenario credit automation modeling method. Figure 1 is a flow chart according to the first embodiment of the present disclosure.

[0036] like Figure 1 As shown, the scenario credit automation modeling method may include but is not limited to the following steps:

[0037] S1: Preprocess the data of multiple new scenario users to obtain a data set; wherein the data set includes a scenario feature set of multiple new scenario users.

[0038] In the disclosed embodiment, preprocessing is performed on the data of multiple new scenario users, which may include checking data quality, missing value processing, single value processing, abnormal value processing, etc.

[0039] Among them, checking data quality can be: 1. Calculating the missing and zero value situations of data fields of users in new scenarios, and eliminating fields where the proportion of users with missing or zero values ​​is more than a certain proportion. For example, eliminating fields where the proportion of users with missing or zero values ​​is more than 50%. 2. Eliminating fields with too concentrated values. For example, calculating the difference between the 90th percentile value and the 10th percentile value of each field and dividing it by the median value, and eliminating fields where the value is less than 5.

[0040] Missing value processing can be to eliminate features with a missing rate or unavailability rate above a certain percentage. Exemplarily, features with a missing rate or unavailability rate above 50% are eliminated. After elimination, for numerical missing features, they are uniformly filled with 0; for missing values of categorical features, they are uniformly filled with 'NotAvailable'.

[0041] Single value processing can be to eliminate features with a single value rate above a certain percentage. Exemplarily, features with a single value rate above 70% are eliminated, or features with a single value rate above 80% are eliminated, etc.

[0042] Outlier processing can be to fill the upper and lower limits, etc.

[0043] It can be understood that each new scenario user may include multiple features. In the embodiments of the present disclosure, all features of the new scenario user are classified into the scenario feature set of the new scenario user, and the scenario feature sets of multiple new scenario users together constitute a data set.

[0044] S2: Determine one or more target existing scenario credit models according to the scenario feature sets of multiple new scenario users and multiple existing scenario credit models.

[0045] It can be understood that in the related art, different scenario credit models have been set for different scenarios. The number of existing scenario credit models is as many as nearly a hundred or more. The existing scenario credit models include the existing scenario credit models within the bank, and can also include the existing scenario credit models in the industry. Exemplarily, the existing scenario credit models include: the tax payment agency credit model and the national tax credit model based on tax data, the existing mortgage credit model based on the stock mortgage, etc., settlement cloud loan credit, merchant cloud loan credit, cloud tax loan credit, salary loan credit, etc. Each credit model has its own access and credit rules and limits.

[0046] In the embodiments of the present disclosure, one or more target existing scenario credit models can be determined from multiple existing scenario credit models according to the scenario feature set of the new scenario user. Among them, one or more target existing scenario credit models can be determined according to the scenario feature set of the new scenario user with certain selection conditions or rules. Exemplarily, one or more existing scenario credit models can be randomly determined, or one or more newly established existing scenario credit models can be selected, or the scenario feature set of the new scenario user is compared with the selected feature set of the existing scenario credit model, and one or more closest existing scenario credit models are selected, etc. The embodiments of the present disclosure do not make specific limitations on this.

[0047] S3: Based on the scenario feature set of the new scenario users, one or more determined target existing scenario credit models, and / or the credit limits of multiple existing users in the one or more determined target existing scenario credit models, obtain the target scenario feature set and target credit limit of the new scenario users.

[0048] In the embodiments of the present disclosure, after determining one or more target existing scenario credit models, based on the scenario feature set of the new scenario users and the one or more determined target existing scenario credit models, obtain the target scenario feature set and target credit limit of the new scenario users; or based on the scenario feature set of the new scenario users, the one or more determined target existing scenario credit models, and the credit limits of multiple existing users in the one or more determined target existing scenario credit models, obtain the target scenario feature set and target credit limit of the new scenario users.

[0049] Among them, the one or more determined target existing scenario credit models have corresponding selected feature sets, and the features in the selected feature sets are the features of existing users. In the case where at least one feature of the selected feature set is included in the scenario feature set of the new scenario users, it indicates that the new scenario users have the features of existing users. In the embodiments of the present disclosure, this part of the new scenario users is called the first type of new scenario users.

[0050] In the case where the scenario feature set of the new scenario users does not include the features in the selected feature sets corresponding to the one or more determined target existing scenario credit models, it can be indicated that the new scenario users do not include the features of existing users. In the embodiments of the present disclosure, this part of the new scenario users is called the second type of new scenario users.

[0051] S4: Based on the target scenario feature sets and target credit limits of multiple new scenario users, construct a new scenario credit model.

[0052] By implementing the embodiments of the present disclosure, preprocess the data of multiple new scenario users to obtain a data set; among them, the data set includes the scenario feature set of each new scenario user; based on the scenario feature set of the new scenario users and multiple existing scenario credit models, determine one or more target existing scenario credit models; based on the scenario feature set of the new scenario users and the one or more determined target existing scenario credit models, obtain the target scenario feature set and target credit limit of the new scenario users; based on the target scenario feature set and target credit limit, construct a new scenario credit model. Thus, it is possible to automatically model the imported new scenario users, greatly improving the efficiency of establishing a credit model, saving labor costs and time costs, and the automatically established credit model can better meet the credit needs of the new credit business scenario.

[0053] In some embodiments, as Figure 2 shown, S2 in the embodiments of the present disclosure includes but is not limited to the following sub-steps:

[0054] S21: Compare the scenario feature set of each new scenario user with the selected feature set of each existing scenario credit model respectively.

[0055] It can be understood that an existing scenario credit model has its corresponding selected feature set, and the selected feature set includes the features of at least one existing user.

[0056] Exemplarily, for the corporate account transaction data corresponding to the settlement cloud loan credit: the average daily balance in the past 12 months, the average daily balance in the past 3 months, the total debit transaction amount in the past 3 months, and the total credit transaction amount in the past 3 months.

[0057] For the personal merchant cloud loan credit corresponding to the personal account transaction data: the maximum balance in the past 6 months, the total transaction amount in the past 3 months.

[0058] For the cloud tax loan credit corresponding to the enterprise tax payment data / national tax payment data: the total tax payment amount in the past 1 month, the total tax payment amount in the past 12 months, and the total tax payment amount in the past 12 to 24 months.

[0059] For the AUM loan credit corresponding to the AUM amount data: the maximum value of all types of AUM in the past 6 months.

[0060] For the salary loan credit corresponding to the enterprise payroll data: the maximum payroll in the past 12 months.

[0061] It should be noted that the features in the selected feature set of the existing scenario credit model are the first type of features, and the first type of features are the features of existing users. The features in the scenario feature set of each new scenario user may or may not be the first type of features. The scenario feature set of each new scenario user may include both the first type of features and the features that are not the first type of features, or may only include the first type of features, or may only include the features that are not the first type of features.

[0062] In the embodiments of the present disclosure, by comparing the features in the scenario feature set of each new scenario user with the features in the selected feature set of each existing scenario credit model respectively, the result of whether the new scenario user covers the existing scenario credit model can be obtained. Further, execute S22.

[0063] S22: Obtain the number of new scenario users who cover each existing scenario credit model; wherein, when the scenario feature set of the new scenario user includes all the features in the selected feature set of the existing scenario credit model and the corresponding features have values within a certain period, it is determined that the new scenario user covers the existing scenario credit model.

[0064] In the disclosed embodiment, the scenario feature set of each new scenario user is compared with the selected feature set of each existing scenario credit model. The scenario feature set of the new scenario user includes all the features in the selected feature set of the existing scenario credit model, and each feature has a value within a certain period of one year, which is considered that the new scenario user covers the existing scenario credit model.

[0065] Among them, having a value within a certain period of time may mean having a value within one year and being greater than 0.

[0066] In the exemplary embodiment, the scenario feature set of a new scenario user includes: the maximum wage payment in the past 12 months is 8,000, the maximum balance in the past 6 months is 28,000, and the total turnover in the past 3 months is 12,000; the selected feature set of the first existing scenario credit model is the maximum wage payment in the past 12 months, and the selected feature set of the second existing scenario credit model is the maximum balance in the past 6 months and the total turnover in the past 3 months. By comparison, it can be seen that the new scenario user covers the first existing scenario credit model and the second existing scenario credit model.

[0067] In another exemplary embodiment, the scenario feature set of a new scenario user includes: the maximum amount of wage payment in the past 12 months is 8,000, and the total amount of turnover in the past 3 months is 12,000; the selected feature set of the first existing scenario credit model is the maximum amount of wage payment in the past 12 months, and the selected feature set of the second existing scenario credit model is the maximum balance in the past 6 months and the total amount of turnover in the past 3 months. By comparison, it can be seen that the new scenario user covers the first existing scenario credit model, but not the second existing scenario credit model.

[0068] It should be noted that the above examples are for illustration only and are not intended to be specific limitations on the embodiments of the present disclosure. The scenario feature set of the scenario user in the embodiments of the present disclosure may also include other features, and the optional feature set of the existing scenario credit model may also include other features. The embodiments of the present disclosure do not impose specific limitations on this.

[0069] S23: Determine an existing scenario credit model with the largest number of users, or, by ranking from most to least in terms of the number of users, determine multiple existing scenario credit models with the highest rankings as the target existing scenario credit model.

[0070] In the embodiment of the present disclosure, an existing scenario credit model with the largest number of users is determined based on the number of new scenario users covered by each existing scenario credit model. Alternatively, the existing scenario credit models are ranked from most to least based on the number of new scenario users covered by each existing scenario credit model, and multiple existing scenario credit models with high rankings are determined as target existing scenario credit models.

[0071] In some embodiments, new scenario users include: the first type of new scenario users and the second type of new scenario users. The scenario feature set of new scenario users includes the first scenario feature set of the first type of new scenario users and the second scenario feature set of the second type of new scenario users. The first scenario feature set includes at least one first type of feature and / or at least one second type of feature. The second scenario feature set does not include the first type of feature and includes at least one second type of feature. Among them, the first type of feature is the feature of existing users, and the second type of feature is not the feature of existing users.

[0072] As Figure 3 shown, S3 in the embodiments of the present disclosure includes but is not limited to the following sub-steps:

[0073] S31: Obtain the first target scenario feature set and the first target amount of the first type of new scenario users according to the first type of features in the first scenario feature set of the first type of new scenario users, one or more determined target existing scenario credit models, and / or the credit amounts of multiple existing users in one or more determined target existing scenario credit models.

[0074] In the embodiments of the present disclosure, the first target scenario feature set and the first target amount of the first type of new scenario users can be obtained according to the first type of features in the first scenario feature set of the first type of new scenario users, one or more determined target existing scenario credit models, and / or the credit amounts of multiple existing users in one or more determined target existing scenario credit models.

[0075] It can be understood that in the case where the first type of features in the first scenario feature set of the first type of new scenario users includes all the features of the selected feature set corresponding to one determined target existing scenario credit model, inputting the first type of features in the first scenario feature set of the first type of new scenario users into one determined target existing scenario credit model can obtain a hypothetical credit amount. In the embodiments of the present disclosure, the hypothetical amount can be used as the first target amount of the first type of new scenario users; alternatively, in the embodiments of the present disclosure, the first type of new scenario users and existing users can also be associated, and the credit amount of the existing users can be referred to to obtain the target amount of the first type of new scenario users, etc.

[0076] In the embodiments of the present disclosure, on the basis of obtaining the first target amount of the first type of new scenario users, the first target scenario feature set can be determined according to the first target amount and the second type of features in the first scenario feature set.

[0077] S32: Obtain the second target scenario feature set and the second target amount of the second type of new scenario users according to the second scenario feature set of the second type of new scenario users and the first target amount of the first type of new scenario users.

[0078] In the embodiments of the present disclosure, the second target scenario feature set and the second target amount of the second type of new scenario users can be obtained based on the second scenario feature set of the second type of new scenario users and the first target amount of the first type of new scenario users.

[0079] In some embodiments, to obtain the first target scenario feature set and the first target amount of the first type of new scenario users according to the first type of features in the first scenario feature set of the first type of new scenario users, one or more determined target existing scenario credit models, and / or the credit amounts of multiple stock users in one or more determined target existing scenario credit models, includes: when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models to obtain one or more assumed scenario credit amounts, determining the first target amount of the first type of new scenario users according to the one or more assumed scenario credit amounts. Calculating the correlation coefficient between the second type of features in the first scenario feature set of the first type of new scenario users and the first target amount, arranging them in descending order according to the correlation coefficient, and selecting the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of selectable features to generate the first target scenario feature set of the first type of new scenario users.

[0080] In some embodiments, when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models to obtain one or more assumed scenario credit amounts, determining the first target amount of the first type of new scenario users includes: when the first type of features in the first scenario feature set of the first type of new scenario users are input into one determined target existing scenario credit model to obtain one assumed scenario credit amount, using the obtained assumed scenario credit amount as the first target amount of the first type of new scenario users, or when the first type of features in the first scenario feature set of the first type of new scenario users are input into multiple determined target existing scenario credit models to obtain multiple assumed scenario credit amounts, using the highest assumed scenario credit amount among the obtained multiple assumed scenario credit amounts as the first target amount of the first type of new scenario users.

[0081] In the embodiments of the present disclosure, the correlation coefficient between the second type of features in the first scenario feature set of the first type of new scenario users and the target amount is calculated. The calculation method of the correlation coefficient is the Pearson correlation coefficient, and the calculation formula is:

[0082]

[0083] where N is the number of samples, x i is the i-th sample value of variable x, is the mean value of variable x, y i is the i-th sample value of variable y, is the mean value of variable y; the correlation coefficient ranges from -1 to 1, and the closer it is to 0, the smaller the correlation.

[0084] In the embodiments of the present disclosure, according to the arrangement from large to small of the correlation coefficients, the second type of features with correlation coefficients greater than the correlation threshold and not exceeding the number of optional features are selected to generate the first target scenario feature set of the first type of new scenario users.

[0085] Exemplarily, in the embodiments of the present disclosure, for the new scenario users among the first type of new scenario users that cover the target existing scenario credit models, calculate the correlation (Pearson distance) between the second type of features in the first scenario feature set and the first target amount, and select the second type of features whose correlation with the first target amount is above 0.3 times the correlation of the most relevant feature and filtered by the self-correlation threshold of 0.8. At the same time, limit the number of features according to the requirement that the ratio of the number of features to the number of samples is greater than 100:1 (when the number of features is greater than one percent of the number of samples, take the maximum number of optional features as the upper limit, and select several second type of features in descending order of correlation).

[0086] In some embodiments, the method further includes: when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models and the assumed scenario credit amount cannot be obtained, determine the first target amount of the first type of new scenario users through the KNN regression algorithm.

[0087] In the embodiments of the present disclosure, the first type of new scenario users include the first type of features, but do not cover all the features of the selected feature set corresponding to any target existing scenario credit model. When the first type of features in the first scenario feature set of this part of the first type of new scenario users include some features of the selected feature set corresponding to one or more determined target existing scenario credit models or do not include the features of the selected feature set corresponding to one or more target existing scenario credit models, inputting the first type of features in the first scenario feature set of the first type of new scenario users into one determined target existing scenario credit model will not obtain the assumed credit amount.

[0088] In this case, in the embodiments of the present disclosure, the first target amount of the first type of new scenario users is determined through the KNN regression algorithm.

[0089] Among them, the KNN regression algorithm can also be called the K-nearest neighbor algorithm. K-nearest neighbor means K nearest neighbors. When predicting an unknown sample, it is determined by the K neighbors closest to the sample. In the embodiments of the present disclosure, the KNN regression algorithm is used. When performing regression prediction, the mean (or weighted mean) of the K neighbors is used as the prediction result. The KNN regression algorithm is easy to understand and can obtain good performance without much adjustment. Unlike other supervised learning methods that require training a model in advance, it has no explicit training process and the training time is zero. It only needs to directly predict the prediction sample through the training sample, greatly improving the model prediction efficiency. In addition, the KNN regression algorithm has no assumptions about the data, has high accuracy, and is not sensitive to outliers, so as to determine the first target amount of the first type of new scenario users more accurately.

[0090] In some embodiments, determining the first target amount of the first type of new scenario users through the KNN regression algorithm includes:

[0091] Obtain multiple existing users who meet the first preset condition in one or more determined target existing scenario credit models, and determine the first target amount of the first type of new scenario users according to the credit amounts of the multiple existing users in the one or more determined target existing scenario credit models.

[0092] In some embodiments of the method, it further includes: in the case of determining one target existing scenario credit model, obtain multiple existing users who meet the first preset condition in the determined one target existing scenario credit model, and take the average value of the multiple credit amounts of the multiple existing users in the determined one target existing scenario credit model as the first target amount of the first type of new scenario users; or, in the case of determining multiple target existing scenario credit models, obtain multiple existing users who meet the first preset condition in each determined target existing scenario credit model, obtain the average value of the multiple credit amounts of the multiple existing users in each target existing scenario credit model, and take the maximum value among the average values of the multiple target existing scenario credit models as the first target amount of the first type of new scenario users.

[0093] In some embodiments, the multiple existing users who meet the first preset condition in the target existing scenario credit model are multiple existing users who have an initial amount in the target existing scenario credit model within a preset duration before the new scenario user import time point and the initial amount is less than the upper limit of the new scenario expected amount.

[0094] It can be understood that the multiple existing users in the target existing scenario credit model that meet the first preset condition can be multiple existing users who had an initial credit limit in the target existing scenario credit model within the first half year before the new scenario user import time point and the initial credit limit is less than the upper limit of the new scenario expected credit limit, or can also be multiple existing users who had an initial credit limit in the target existing scenario credit model within 1 year before the new scenario user import time point and the initial credit limit is less than the upper limit of the new scenario expected credit limit. The embodiments of the present disclosure do not make specific restrictions on this.

[0095] In some embodiments, as Figure 4 shown, S32 in the embodiments of the present disclosure includes but is not limited to the following sub-steps:

[0096] S321: Determine the second target credit limit of the second type of new scenario users through the KNN regression algorithm according to the second scenario feature set of the second type of new scenario users and the first target credit limit of the first type of new scenario users.

[0097] In some embodiments, determining the second target credit limit of the second type of new scenario users through the KNN regression algorithm according to the second scenario feature set of the second type of new scenario users and the first target credit limit of the first type of new scenario users includes: Selecting multiple first type of new scenario users that meet the second preset condition from the first type of new scenario users according to the second scenario feature set of the second type of new scenario users, and obtaining the average value of the first target credit limits of the selected multiple first type of new scenario users as the second target credit limit of the second type of new scenario users.

[0098] S322: Calculate the correlation coefficient between the second type of features in the second scenario feature set of the second type of new scenario users and the target credit limit, arrange them from largest to smallest according to the correlation coefficient, select the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of selectable features, and generate the second target scenario feature set of the second type of new scenario users.

[0099] In some embodiments, as Figure 5 shown, S4 in the embodiments of the present disclosure includes but is not limited to the following sub-steps:

[0100] S41: Use the target scenario feature sets of multiple new scenario users as independent variables and the corresponding target credit limits as dependent variables to generate a training sample data set.

[0101] In the embodiments of the present disclosure, through the above steps, the target scenario feature set and the target credit limit of each new scenario user can be obtained. The target scenario feature sets and the target credit limits of each new scenario user are put into one-to-one correspondence, and used as independent variables and dependent variables respectively, and then summarized to generate a training sample data set.

[0102] S42: Input the training sample data set into the LightGBM model, train the LightGBM model, and obtain a new scenario credit model.

[0103] Among them, the LightGBM model is an integrated learning model with the nonlinear model GBDT (Gradient Boosting DecisionTree) algorithm as the core, decision tree as the base classifier, and layer growth tree as the strategy. It aims to gradually improve the base learner through the iterative process of multiple decision trees until the number of base learners reaches the target value. As the most advanced integrated learning model framework, Light GBM has the characteristics of direct support for category features, multi-threaded optimization, and support for efficient parallelism compared to other integrated models. In addition, for the construction of credit models, the use of advanced and complex machine learning technology can make more full use of data and more comprehensive and effective identification of complex risks.

[0104] In the disclosed embodiment, the constructed new scenario credit model also includes adjusting the credit limit. For example, the credit limit is adjusted based on the nature of the enterprise, industry and risk rating. In the disclosed embodiment, the credit limit can be adjusted considering the user's risk, industry, nature of the enterprise and other information. Based on the existing stock users, the relationship between the user's credit limit and basic information is summarized, and the credit limit adjustment coefficients including user risk, user industry, nature of the user's enterprise and so on are added.

[0105] Among them, enterprise risk adjustment includes: based on the risk score card, the existing users are divided into boxes according to the score card PDO value as the box span, and the median credit limit of users in each box is divided by the median credit limit of all existing users as the credit limit adjustment coefficient of each box.

[0106] After the initial credit limit is calculated for new scenario users through the new scenario credit model, the scoring card for the new scenario user is called, and the adjustment coefficient of the subbox to which the risk score of the new scenario user belongs is used as the risk coefficient of the new scenario user. The initial credit limit is multiplied by the risk coefficient as the risk-adjusted credit limit.

[0107] Industry adjustment includes: Based on industry categories, classifying existing users, and obtaining an adjustment ratio for each category by dividing the median amount of users within each industry category by the median amount of all existing users. On this basis, combined with the current macro-policy situation of each industry (refined to the fourth-level industry), for industries where the current macro-policy requires reduction, if the adjustment ratio of the category it belongs to is greater than 1, then multiply the adjustment ratio by 0.8 as the adjustment coefficient (if it becomes less than 1 after multiplication, set the adjustment coefficient to 1); for industries where the current macro-policy requires support, if the adjustment ratio of the category it belongs to is less than 1, then multiply the adjustment ratio by 1.2 as the adjustment coefficient (if it becomes greater than 1 after multiplication, set the adjustment coefficient to 1); for those without a clear policy orientation, use the adjustment ratio of the category it belongs to as the adjustment coefficient.

[0108] For new-scenario users, the amount output after the previous risk adjustment is multiplied by the industry adjustment coefficient of the fourth-level industry they belong to as the amount used in this link.

[0109] Enterprise nature adjustment includes: Based on enterprise nature, classifying existing users, and obtaining an adjustment coefficient for each enterprise nature category by dividing the median amount of users in each enterprise nature category by the median amount of all existing users. For new-scenario users, the amount output after the previous industry adjustment is multiplied by the adjustment coefficient of the enterprise nature category they belong to as the final output amount.

[0110] In the embodiments of the present disclosure, after constructing the new-scenario credit-granting model, the constructed new-scenario credit-granting model will also be monitored and iterated.

[0111] Among them, model monitoring: After the new-scenario credit-granting model is launched, a corresponding monitoring plan needs to be established, and the branch or the head office monitors the model data and model results to understand the operation of the new-scenario credit-granting model and the business change trend, ensuring that the new-scenario credit-granting model operates stably in the business scenario.

[0112] Model monitoring mainly includes two aspects: On the one hand, monitor the data used by the new-scenario credit-granting model and its own performance, such as quality, stability, and anomaly monitoring; on the other hand, monitor business-level information, such as operation, call times, and amount distribution, etc.

[0113] Exemplarily, as shown in Table 1 below, in the embodiments of the present disclosure, the model is monitored.

[0114] Table 1:

[0115]

[0116]

[0117] In an embodiment of the present disclosure, after constructing a new scenario credit model, the constructed new scenario credit model is iterated.

[0118] After the new scenario credit model is launched, model iteration can be performed according to the situation of the monitoring report or business requirements. Over time, due to external business factors or requirements, abnormal distribution of imported data in the same scenario, abnormal stability of the model itself, etc., the model effect will decline accordingly. At this time, the model needs to be iterated, that is, reselect the imported sample data or rebuild the model, and the new model re-predicts the quota.

[0119] Figure 6 It is a structural diagram according to the second embodiment of the present disclosure.

[0120] Such as Figure 6 As shown, the automated scenario credit modeling device 10 includes a data set acquisition unit 11, a model selection unit 12, a processing unit 13, and a model construction unit 14, where:

[0121] The data set acquisition unit 11 is configured to preprocess the data of multiple new scenario users to obtain a data set; wherein, the data set includes scenario feature sets of multiple new scenario users.

[0122] The model selection unit 12 is configured to determine one or more target existing scenario credit models according to the scenario feature sets of multiple new scenario users and multiple existing scenario credit models.

[0123] The processing unit 13 is configured to obtain the target scenario feature set and target quota of the new scenario user according to the scenario feature set of the new scenario user, the determined one or more target existing scenario credit models, and / or the credit quotas of multiple existing users in the determined one or more target existing scenario credit models.

[0124] The model construction unit 14 is configured to construct a new scenario credit model according to the target scenario feature sets and target quotas of multiple new scenario users.

[0125] By implementing the embodiments of the present disclosure, the dataset acquisition unit 11 is configured to preprocess the data of multiple new-scenario users to obtain a dataset; wherein the dataset includes the scenario feature sets of multiple new-scenario users; the model selection unit 12 is configured to determine one or more target existing-scenario credit models according to the scenario feature sets of multiple new-scenario users and multiple existing-scenario credit models; the processing unit 13 is configured to obtain the target scenario feature set and target quota of new-scenario users according to the scenario feature sets of new-scenario users, the determined one or more target existing-scenario credit models, and / or the credit quotas of multiple stock users in the determined one or more target existing-scenario credit models; the model construction unit 14 is configured to construct a new-scenario credit model according to the target scenario feature sets and target quotas of multiple new-scenario users. Thus, it is possible to automatically model the imported new-scenario users, greatly improving the efficiency of establishing a credit model, saving labor costs and time costs, and the automatically established credit model can better meet the credit needs of new credit business scenarios.

[0126] In some embodiments, as Figure 7 shown, in the embodiments of the present disclosure, the model selection unit 12 includes:

[0127] The dataset comparison subunit 121 is configured to compare the scenario feature set of each new-scenario user with the selected feature set of each existing-scenario credit model respectively.

[0128] The user quantity calculation subunit 122 is configured to obtain the quantity of new-scenario users covering each existing-scenario credit model; wherein, when the scenario feature set of a new-scenario user includes all the features in the selected feature set of an existing-scenario credit model and the corresponding features have values within a certain period, it is determined that the new-scenario user covers the existing-scenario credit model.

[0129] The model selection subunit 123 is configured to determine an existing-scenario credit model with the largest user quantity, or determine multiple top-ranked existing-scenario credit models ranked from more to less user quantity as the target existing-scenario credit models.

[0130] In some embodiments, as Figure 8 shown, in the embodiments of the present disclosure, new-scenario users include: the first type of new-scenario users and the second type of new-scenario users. The scenario feature set of new-scenario users includes the first scenario feature set of the first type of new-scenario users and the second scenario feature set of the second type of new-scenario users. The first scenario feature set includes at least one first type of feature and / or at least one second type of feature, and the second scenario feature set does not include the first type of feature and includes at least one second type of feature. Wherein, the first type of feature is the feature of stock users, and the second type of feature is not the feature of stock users; the processing unit 13 includes:

[0131] The first processing subunit 131 is configured to obtain a first target scenario feature set and a first target quota of the first type of new scenario users according to the first type of features in the first scenario feature set of the first type of new scenario users, one or more determined target existing scenario credit models, and / or the credit quotas of multiple stock users in the one or more determined target existing scenario credit models.

[0132] The second processing subunit 132 is configured to obtain a second target scenario feature set and a second target quota of the second type of new scenario users according to the second scenario feature set of the second type of new scenario users and the first target quota of the first type of new scenario users.

[0133] In some embodiments, as Figure 9 shown, in the embodiments of the present disclosure, the first processing subunit 131 includes:

[0134] The first target quota determination module 1311 is configured to determine the first target quota of the first type of new scenario users according to one or more assumed scenario credit quotas when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models to obtain one or more assumed scenario credit quotas.

[0135] The first target scenario feature set determination module 1312 is configured to calculate the correlation coefficients between the second type of features in the first scenario feature set of the first type of new scenario users and the target quota, select the second type of features with correlation coefficients greater than the correlation threshold and not exceeding the number of selectable features according to the descending order of the correlation coefficients, and generate the first target scenario feature set of the first type of new scenario users.

[0136] The second target quota determination module 1313 is configured to determine the first target quota of the first type of new scenario users through the KNN regression algorithm when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models and no assumed scenario credit quota can be obtained.

[0137] In some embodiments, the first target quota determination module 1311 is specifically configured to use the obtained assumed scenario credit quota as the first target quota of the first type of new scenario users when the first type of features in the first scenario feature set of the first type of new scenario users are input into one determined target existing scenario credit model to obtain one assumed scenario credit quota, or use the highest assumed scenario credit quota among the obtained multiple assumed scenario credit quotas as the first target quota of the first type of new scenario users when the first type of features in the first scenario feature set of the first type of new scenario users are input into multiple determined target existing scenario credit models to obtain multiple assumed scenario credit quotas.

[0138] In some embodiments, the second target quota determination module 1313 is specifically configured to obtain multiple existing users who meet the first preset condition in one or more determined target existing scenario credit models, and determine the first target quota of the first type of new scenario users according to the credit quotas of the multiple existing users in the one or more determined target existing scenario credit models.

[0139] In some embodiments, the second target quota determination module 1313 is specifically configured to, in the case of determining one target existing scenario credit model, obtain multiple existing users who meet the first preset condition in the determined one target existing scenario credit model, and take the average value of the multiple credit quotas as the first target quota of the first type of new scenario users according to the multiple credit quotas of the multiple existing users in the determined one target existing scenario credit model; in the case of determining multiple target existing scenario credit models, obtain multiple existing users who meet the first preset condition in each determined target existing scenario credit model, obtain the average value of the multiple credit quotas of the multiple existing users in each target existing scenario credit model, and take the maximum value among the average values of the multiple target existing scenario credit models as the first target quota of the first type of new scenario users.

[0140] In some embodiments, among the multiple existing users who meet the first preset condition in the target existing scenario credit model in the second target quota determination module 1313 are multiple existing users who have an initial quota in the target existing scenario credit model within a preset duration before the new scenario user import time point and the initial quota is less than the upper limit of the new scenario expected quota.

[0141] In some embodiments, as Figure 10 shown, in the embodiments of the present disclosure, the second processing subunit 132 includes:

[0142] A third target quota determination module 1321, configured to determine the second target quota of the second type of new scenario users through a KNN regression algorithm according to the second scenario feature set of the second type of new scenario users and the first target quota of the first type of new scenario users.

[0143] A second target scenario feature set determination module 1322, configured to calculate the correlation coefficient between the second type of features in the second scenario feature set of the second type of new scenario users and the target quota, and select the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of selectable features according to the descending order of the correlation coefficients to generate the second target scenario feature set of the second type of new scenario users.

[0144] In some embodiments, the third target amount determination module 1321 is specifically configured to select, according to the second scenario feature set of the second type of new scenario users, multiple first type of new scenario users that meet the second preset condition among the first type of new scenario users, and obtain the average value of the first target amounts of the selected multiple first type of new scenario users as the second target amount of the second type of new scenario users.

[0145] In some embodiments, as Figure 11 shown, in the embodiments of the present disclosure, the model construction unit 14 includes:

[0146] A training data set acquisition subunit 141, configured to generate a training sample data set by using the target scenario feature sets of multiple new scenario users as independent variables and the corresponding target amounts as dependent variables;

[0147] A model training and construction subunit 142, configured to input the training sample data set into a LightGBM model, train the LightGBM model, and obtain a new scenario credit granting model.

[0148] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0149] To implement the above embodiments, the embodiments of the present disclosure also propose an electronic device. The electronic device includes: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the scenario credit granting automated modeling method as described above.

[0150] As an example, Figure 12 is a block diagram of an electronic device 200 for a scenario credit granting automated modeling method shown according to an exemplary embodiment. As Figure 12 shown, the above electronic device 200 may further include:

[0151] A memory 210 and a processor 220, a bus 230 connecting different components (including the memory 210 and the processor 220), and the memory 210 stores a computer program. When the processor 220 executes the program, the scenario credit granting automated modeling method of the embodiments of the present disclosure is implemented.

[0152] Bus 230 represents one or more of several types of bus architectures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. By way of example, and not limitation, these architectures include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0153] Electronic device 200 typically includes a variety of electronic device-readable media. These media can be any available media that can be accessed by electronic device 200, including both volatile and nonvolatile media, removable and non-removable media.

[0154] Memory 210 may also include computer system-readable media in the form of volatile memory, such as random access memory (RAM) 240 and / or cache memory 250. Server 200 may further include other removable / non-removable, volatile / nonvolatile computer system storage media. By way of example only, storage system 260 can be used for reading from and writing to non-removable, nonvolatile magnetic media ( Figure 12 not shown, and typically called a "hard disk drive"). Although Figure 12 not shown in the figures, a disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from and writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) can be provided. In these instances, each drive can be connected to bus 230 by one or more data media interfaces. Memory 210 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the present disclosure.

[0155] A program / utility 280 having a set (at least one) of program modules 270 can be stored, for example, in memory 210, and such program modules 270 include—but are not limited to—an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may include an implementation of a network environment. Program modules 270 generally carry out the functions and / or methods of the embodiments described herein.

[0156] The electronic device 200 can also communicate with one or more external devices 290 (such as a keyboard, a pointing device, a display 291, etc.), and can also communicate with one or more devices that enable a user to interact with the electronic device 200, and / or communicate with any device that enables the electronic device 200 to communicate with one or more other computing devices (such as a network card, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 292. Moreover, the electronic device 200 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 293. As shown in the figure, the network adapter 293 communicates with other modules of the electronic device 200 through a bus 230. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0157] The processor 220 executes various functional applications and data processing by running programs stored in the memory 210.

[0158] It should be noted that for the implementation process and technical principle of the electronic device in this embodiment, refer to the foregoing explanation of the scenario credit automation modeling method of the embodiments of the present disclosure, and details will not be elaborated here.

[0159] To implement the above embodiments, the embodiments of the present disclosure also propose a storage medium.

[0160] Wherein, when the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute the scenario credit automation modeling method as described above.

[0161] To implement the above embodiments, the present disclosure also provides a computer program product. When the computer program is executed by the processor of the electronic device, the electronic device can execute the scenario credit automation modeling method as described above.

[0162] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0163] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A method for automated modeling of scenario credit granting, characterized in that, Including: Preprocessing the data of multiple new-scenario users to obtain a data set; wherein, the data set includes the scenario feature sets of multiple new-scenario users; Determining one or more target existing-scenario credit models according to the scenario feature sets of multiple new-scenario users and multiple existing-scenario credit models; Obtaining the target scenario feature set and target amount of the new-scenario users according to the scenario feature set of the new-scenario users, the determined one or more target existing-scenario credit models, and / or the credit amounts of multiple existing users in the determined one or more target existing-scenario credit models; Constructing a new-scenario credit model according to the target scenario feature set and the target amount of multiple new-scenario users.

2. The method according to claim 1, wherein The determining one or more target existing-scenario credit models according to the scenario feature sets of multiple new-scenario users and multiple existing-scenario credit models includes: Comparing the scenario feature set of each new-scenario user with the selected feature set of each existing-scenario credit model respectively; Obtaining the number of new-scenario users covering each existing-scenario credit model; wherein, when the scenario feature set of the new-scenario user includes all the features in the selected feature set of the existing-scenario credit model and all the features have values within one year, it is determined that the new-scenario user covers the existing-scenario credit model; Determining an existing-scenario credit model with the largest number of users, or determining multiple top-ranked existing-scenario credit models according to the number of users from more to less as the target existing-scenario credit models.

3. The method according to claim 2, wherein The new-scenario users include: the first type of new-scenario users and the second type of new-scenario users. The scenario feature set of the new-scenario users includes the first scenario feature set of the first type of new-scenario users and the second scenario feature set of the second type of new-scenario users. The first scenario feature set includes at least one first type of feature and at least one second type of feature. The second scenario feature set does not include the first type of feature and includes at least one second type of feature. Among them, the first type of feature is the feature of existing users, and the second type of feature is not the feature of existing users. The obtaining the target scenario feature set and target amount of the new-scenario users according to the scenario feature set of the new-scenario users, the determined one or more target existing-scenario credit models, and / or the credit amounts of multiple existing users in the determined one or more target existing-scenario credit models includes: Obtaining the first target scenario feature set and the first target amount of the first type of new-scenario users according to the first type of features in the first scenario feature set of the first type of new-scenario users, the determined one or more target existing-scenario credit models, and / or the credit amounts of multiple existing users in the determined one or more target existing-scenario credit models; Obtaining the second target scenario feature set and the second target amount of the second type of new-scenario users according to the second scenario feature set of the second type of new-scenario users and the first target amount of the first type of new-scenario users.

4. The method according to claim 3, characterized in that One or more target existing scenario credit models determined according to the first type of features in the first scenario feature set of the first type of new scenario users, and / or the credit limits of multiple existing users in one or more target existing scenario credit models determined, obtaining the first target scenario feature set and the first target limit of the first type of new scenario users, includes: When the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more target existing scenario credit models determined, and one or more assumed scenario credit limits are obtained, determining the first target limit of the first type of new scenario users according to the one or more assumed scenario credit limits; Calculating the correlation coefficient between the second type of features in the first scenario feature set of the first type of new scenario users and the first target limit, arranging according to the correlation coefficient from large to small, selecting the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of selectable features, and generating the first target scenario feature set of the first type of new scenario users.

5. The method according to claim 4, wherein The determining the first target limit of the first type of new scenario users according to the one or more assumed scenario credit limits when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more target existing scenario credit models determined and one or more assumed scenario credit limits are obtained, includes: When the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more target existing scenario credit models determined and one assumed scenario credit limit is obtained, taking the obtained assumed scenario credit limit as the first target limit of the first type of new scenario users; Or, When the first type of features in the first scenario feature set of the first type of new scenario users are input into multiple target existing scenario credit models determined and multiple assumed scenario credit limits are obtained, taking the highest assumed scenario credit limit among the obtained multiple assumed scenario credit limits as the first target limit of the first type of new scenario users.

6. The method according to claim 4 or 5, characterized in that, The method further includes: When the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more target existing scenario credit models determined and no assumed scenario credit limit can be obtained, determining the first target limit of the first type of new scenario users through the KNN regression algorithm.

7. The method according to claim 6, wherein The determining the first target limit of the first type of new scenario users through the KNN regression algorithm includes: Obtaining multiple existing users that meet the first preset condition in one or more target existing scenario credit models determined, and determining the first target limit of the first type of new scenario users according to the credit limits of the multiple existing users in one or more target existing scenario credit models determined.

8. The method according to claim 7, wherein The method further includes: When a target existing scenario credit model is determined, obtain multiple existing users in the determined target existing scenario credit model who meet the first preset condition. According to the multiple credit limits of the multiple existing users in the determined target existing scenario credit model, take the average value of the multiple credit limits as the first target limit of the first type of new scenario users; Or, When multiple target existing scenario credit models are determined, obtain multiple existing users in each determined target existing scenario credit model who meet the first preset condition, obtain the average value of the multiple credit limits of the multiple existing users in each target existing scenario credit model, and take the maximum value among the average values of the multiple existing scenario credit models as the first target limit of the first type of new scenario users.

9. The method according to claim 7 or 8, characterized in that, The multiple existing users in the target existing scenario credit model who meet the first preset condition are multiple existing users who have an initial limit in the target existing scenario credit model within a preset duration before the new scenario user import time point, and the initial limit is less than the upper limit of the new scenario expected limit.

10. The method according to claim 3, wherein The obtaining of the second target scenario feature set and the second target limit of the second type of new scenario users according to the second scenario feature set of the second type of new scenario users and the first target limit of the first type of new scenario users includes: According to the second scenario feature set of the second type of new scenario users and the first target limit of the first type of new scenario users, determine the second target limit of the second type of new scenario users through the KNN regression algorithm; Calculate the correlation coefficient between the second type of features in the second scenario feature set of the second type of new scenario users and the second target limit, arrange them from large to small according to the correlation coefficient, and select the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of optional features to generate the second target scenario feature set of the second type of new scenario users.

11. The method according to claim 10, characterized in that, The determining of the second target limit of the second type of new scenario users through the KNN regression algorithm according to the second scenario feature set of the second type of new scenario users and the first target limit of the first type of new scenario users includes: According to the second scenario feature set of the second type of new scenario users, select multiple first type of new scenario users who meet the second preset condition among the first type of new scenario users, and obtain the average value of the first target limits of the selected multiple first type of new scenario users as the second target limit of the second type of new scenario users.

12. The method according to claim 1, characterized in that, The constructing of a new scenario credit model according to the target scenario feature sets and the target limits of multiple new scenario users includes: Take the target scenario feature sets of multiple new scenario users as independent variables and the corresponding target limits as dependent variables to generate a training sample data set; Input the training sample data set into the LightGBM model and train the LightGBM model to obtain a new scenario credit model.

13. An automated modeling device for scenario credit granting, characterized in that Including: A data set acquisition unit for preprocessing the data of multiple new scenario users to obtain a data set; wherein, the data set includes the scenario feature sets of multiple new scenario users; A model selection unit, configured to determine one or more target existing scenario credit models according to the scenario feature sets of multiple new scenario users and multiple existing scenario credit models; A processing unit, configured to obtain the target scenario feature set and target quota of the new scenario user according to the scenario feature set of the new scenario user, the determined one or more target existing scenario credit models, and / or the credit quotas of multiple existing users in the determined one or more target existing scenario credit models; A model construction unit, configured to construct a new scenario credit model according to the target scenario feature set and the target quota of multiple new scenario users; 14. The device according to claim 13, characterized in that The model selection unit includes: A data set comparison subunit, configured to compare the scenario feature set of each new scenario user with the selected feature set of each existing scenario credit model respectively; A user quantity calculation subunit, configured to obtain the number of new scenario users covering each existing scenario credit model; wherein, when the scenario feature set of the new scenario user includes all the features in the selected feature set of the existing scenario credit model and all the features have values within one year, it is determined that the new scenario user covers the existing scenario credit model; A model selection subunit, configured to determine an existing scenario credit model with the largest number of users, or determine multiple top-ranked existing scenario credit models according to the number of users from more to less as the target existing scenario credit models; 15. The device according to claim 14, characterized in that, The new scenario users include: first-type new scenario users and second-type new scenario users. The scenario feature set of the new scenario user includes the first scenario feature set of the first-type new scenario user and the second scenario feature set of the second-type new scenario user. The first scenario feature set includes at least one first-type feature and at least one second-type feature. The second scenario feature set does not include the first-type feature and includes at least one second-type feature. Among them, the first-type feature is a feature of the existing user, and the second-type feature is not a feature of the existing user; the processing unit includes: A first processing subunit, configured to obtain the first target scenario feature set and the first target quota of the first-type new scenario user according to the first-type features in the first scenario feature set of the first-type new scenario user and the determined one or more target existing scenario credit models; A second processing subunit, configured to obtain the second target scenario feature set and the second target quota of the second-type new scenario user according to the second scenario feature set of the second-type new scenario user and the first target quota of the first-type new scenario user; 16. The device according to claim 15, characterized in that, The first processing subunit includes: A first target quota determination module, configured to, when the first-type features in the first scenario feature set of the first-type new scenario user are input into the determined one or more target existing scenario credit models to obtain one or more assumed scenario credit quotas, determine the first target quota of the first-type new scenario user according to the one or more assumed scenario credit quotas; The first target scenario feature set determination module is configured to calculate the correlation coefficient between the second type of features in the first scenario feature set of the first type of new scenario users and the first target amount, arrange them in descending order according to the correlation coefficient, select the second type of features whose correlation coefficient is greater than the correlation threshold and does not exceed the number of optional features, and generate the first target scenario feature set of the first type of new scenario users.

17. The device according to claim 16, characterized in that, The first target amount determination module is specifically configured to, when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models to obtain a hypothetical scenario credit amount, use the obtained hypothetical scenario credit amount as the first target amount of the first type of new scenario users; or, when the first type of features in the first scenario feature set of the first type of new scenario users are input into multiple determined target existing scenario credit models to obtain multiple hypothetical scenario credit amounts, use the highest hypothetical scenario credit amount among the obtained multiple hypothetical scenario credit amounts as the first target amount of the first type of new scenario users.

18. The device according to claim 16 or 17, characterized in that, The first processing subunit further includes: The second target amount determination module is configured to, when the first type of features in the first scenario feature set of the first type of new scenario users are input into one or more determined target existing scenario credit models and no hypothetical scenario credit amount can be obtained, determine the first target amount of the first type of new scenario users through the KNN regression algorithm.

19. The device according to claim 18, characterized in that, The second target amount determination module is specifically configured to obtain multiple existing users that meet the first preset condition in one or more determined target existing scenario credit models, and determine the first target amount of the first type of new scenario users according to the credit amounts of the multiple existing users in one or more determined target existing scenario credit models.

20. The device according to claim 19, wherein The second target amount determination module is specifically configured to, when one target existing scenario credit model is determined, obtain multiple existing users that meet the first preset condition in the determined one target existing scenario credit model, and take the average value of the multiple credit amounts of the multiple existing users in the determined one target existing scenario credit model as the first target amount of the first type of new scenario users; When multiple target existing scenario credit models are determined, obtain multiple existing users that meet the first preset condition in each determined target existing scenario credit model, obtain the average value of the multiple credit amounts of the multiple existing users in each target existing scenario credit model, and take the maximum value among the average values of the multiple target existing scenario credit models as the first target amount of the first type of new scenario users.

21. The device according to claim 19 or 20, characterized in that, In the second target amount determination module, the multiple existing users that meet the first preset condition in the target existing scenario credit model are multiple existing users who have an initial amount in the target existing scenario credit model within a preset duration before the new scenario user import time point, and the initial amount is less than the upper limit of the new scenario expected amount.

22. The device according to claim 15, characterized in that, The second processing subunit includes: A third target amount determination module, configured to determine a second target amount of the second type of new scenario user through a KNN regression algorithm according to the second scenario feature set of the second type of new scenario user and the first target amount of the first type of new scenario user; A second target scenario feature set determination module, configured to calculate a correlation coefficient between a second type of feature in the second scenario feature set of the second type of new scenario user and the second target amount, arrange them in descending order according to the correlation coefficient, select second type of features whose correlation coefficient is greater than a correlation threshold and does not exceed the number of selectable features, and generate a second target scenario feature set of the second type of new scenario user.

23. The device according to claim 22, characterized in that, The third target amount determination module is specifically configured to, according to the second scenario feature set of the second type of new scenario user, select multiple first type of new scenario users that meet a second preset condition among the first type of new scenario users, and obtain an average value of the first target amounts of the selected multiple first type of new scenario users as the second target amount of the second type of new scenario user.

24. The device according to claim 13, characterized in that, The model construction unit includes: A training data set acquisition subunit, configured to generate a training sample data set by using the target scenario feature sets of multiple new scenario users as independent variables and the corresponding target amounts as dependent variables; A model training and construction subunit, configured to input the training sample data set into a LightGBM model, train the LightGBM model, and obtain a new scenario credit granting model.

25. An electronic device, characterized in that, It includes: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to execute the instructions to implement the method according to any one of claims 1 to 12.

26. A storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, enabling the electronic device to execute the method according to any one of claims 1 to 12.

27. A computer program product, comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the method according to any one of claims 1 to 12.

Images