User credit card limit determination method and device

Abstract

The invention provides a user credit card quota determination method, which can be applied to the technical field of artificial intelligence. The user credit card limit determination method comprises the steps of determining whether a consumption scene of a user belongs to a consumption demand intensive scene according to consumption and browsing contents of the user; under the condition that the consumption scene of the user belongs to the consumption demand intensive scene, obtaining user information of the user, determining a first feature, a second feature and a third feature according to the user information of the user, inputting a user quota-increasing model, and determining a consumption potential score, a risk level score and a scene sensitivity score of the user; the user rating model determines a user comprehensive score according to the consumption potential score, the risk level score and the scene sensitivity score, and determines whether to perform rating on the user according to the user comprehensive score; and under the condition that the quota is determined to be extracted from the user, determining the quota of the user after the quota is extracted. The invention further provides a user credit card quota determination device and equipment, a storage medium and a program product.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a user credit card limit determination method and device. BACKGROUND

[0002] In the prior art, when dividing users, the dimensions relied on are few, and it is difficult to accurately reflect the real value and demand of the users, so that it is not accurate and timely to determine whether to increase the credit limit of the user.

[0003] Therefore, a new user credit card limit determination method and device are needed. SUMMARY

[0004] In view of the above problems, the present application provides a user credit card limit determination method and device.

[0005] According to a first aspect of the present application, a user credit card limit determination method is provided, characterized in that the method comprises: determining whether the consumption scenario of a user belongs to a consumption demand intensive scenario according to the consumption and browsing content of the user, the consumption demand intensive scenario comprising at least one of marriage, car purchase, house purchase, travel, hospitalization, and school; in the case that the consumption scenario of the user belongs to the consumption demand intensive scenario, obtaining user information of the user, and determining a first feature representing merchant preferences of the user, a second feature representing a transaction behavior sequence of the user, and a third feature representing a consumption scenario of the user according to the user information of the user; inputting the first feature, the second feature, and the third feature into a user limit increase model to determine a consumption potential score, a risk level score, and a scenario sensitivity score of the user; the user limit increase model determines a comprehensive score of the user according to the consumption potential score, the risk level score, and the scenario sensitivity score, and determines whether to increase the credit limit of the user according to the comprehensive score of the user; in the case that it is determined to increase the credit limit of the user, determining the credit limit of the user after the limit increase according to at least one of the consumption potential score, the risk level score, and the scenario sensitivity score.

[0006] According to an embodiment of the present application, the user information comprises credit card transaction records, and determining the first feature representing the merchant preferences of the user according to the user information of the user comprises: determining a merchant type according to a merchant name in the credit card transaction records; and determining the first feature according to the merchant type.

[0007] According to an embodiment of the present application, determining the second feature representing the transaction behavior sequence of the user according to the user information of the user comprises: determining target transaction records with transaction continuity according to transaction time and transaction type in the credit card transaction records; and determining the second feature according to the target transaction records.

[0008] According to an embodiment of this application, determining a third feature representing a user's consumption scenario based on the user's user information includes: determining a scenario-based tag for each credit card transaction based on credit card transaction records, wherein the scenario-based tag identifies the user's consumption scenario; and determining the third feature based on the scenario-based tags corresponding to the credit card transactions included in the credit card transaction records.

[0009] According to an embodiment of this application, obtaining user information includes: determining the size of a time window based on the frequency of the user's operation on the online banking application or the frequency of the user's transactions; the higher the frequency of the user's operation on the online banking application or the frequency of the user's transactions, the smaller the time window; determining the time range for obtaining user information based on the size of the time window; and obtaining user information based on the time range.

[0010] According to an embodiment of this application, inputting the first feature, the second feature, and the third feature into a user credit limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score includes: determining the user's consumption potential score for different merchant types based on the first feature and the second feature; determining the user's risk level score based on the second feature and the user's credit card repayment history; and determining the user's scenario sensitivity score based on the first feature and the third feature, wherein the scenario sensitivity score describes the user's willingness to consume in different consumption scenarios.

[0011] According to a second aspect of this application, a user credit card limit determination device is provided, characterized in that the device comprises: a consumption demand intensive scenario identification module, used to determine whether a user's consumption scenario belongs to a consumption demand intensive scenario based on the user's consumption and browsing content, wherein the consumption demand intensive scenario includes at least one of marriage, car purchase, house purchase, travel, hospitalization, and schooling; an initial feature extraction module, used to obtain user information when the user's consumption scenario belongs to a consumption demand intensive scenario, and to determine a first feature representing the user's merchant preference, a second feature representing the user's transaction behavior sequence, and a third feature representing the user's consumption scenario based on the user information; a score calculation module, used to input the first feature, the second feature, and the third feature into a user limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score; a comprehensive score determination module, used to call the user limit increase model to determine the user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and to determine whether to increase the user's credit limit based on the user's comprehensive score; and a credit card limit adjustment module, used to determine the user's increased credit limit based on at least one of the consumption potential score, risk level score, and scenario sensitivity score when it is determined that the user's credit limit will be increased.

[0012] A third aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0013] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0014] The fifth aspect of this application also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method. Attached Figure Description

[0015] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0016] Figure 1 The illustration shows an application scenario of a user credit card limit determination method, apparatus, device, medium, and program product according to embodiments of this application.

[0017] Figure 2 A flowchart illustrating a method for determining a user's credit card limit according to an embodiment of this application is shown schematically.

[0018] Figure 3 This schematic diagram illustrates a structural block diagram of a user credit card limit determination device according to an embodiment of this application;

[0019] Figure 4 A block diagram schematically illustrates an electronic device suitable for implementing a user credit card limit determination method according to an embodiment of this application. Detailed Implementation

[0020] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0021] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0022] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0023] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0024] It should be noted that the user credit card limit determination method, device, equipment, medium, and program products defined in this application can be used in the fields of artificial intelligence technology and fintech, and can also be used in various other fields besides those mentioned above. The application fields of the user credit card limit determination method, device, equipment, medium, and program products provided in the embodiments of this application are not limited.

[0025] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0026] In the technical solution of this application, the user information (including but not limited to user data, user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, and necessary measures have been taken to ensure that they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse.

[0027] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this application all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0028] Embodiments of this application provide a method, apparatus, device, medium, and program product for determining a user's credit card limit.

[0029] Figure 1 The illustration shows an application scenario diagram of the user credit card limit determination method, apparatus, device, medium, and program product according to embodiments of this application.

[0030] like Figure 1 As shown, application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0031] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0032] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0033] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0034] It should be noted that the user credit card limit determination method provided in this application embodiment can generally be executed by server 105. Correspondingly, the user credit card limit determination device provided in this application embodiment can generally be located in server 105. The user credit card limit determination method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the user credit card limit determination device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0035] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0036] The following will be based on Figure 1 The described scene, through Figure 2 The method for determining a user's credit card limit according to an embodiment of this application will be described in detail.

[0037] Figure 2 A flowchart illustrating a method for determining a user's credit card limit according to an embodiment of this application is shown.

[0038] like Figure 2 As shown, the user credit card limit determination method in this embodiment includes steps 210-240, and the user credit card limit determination method can be executed in an electronic device.

[0039] First, perform step 210: Determine whether the user's consumption scenario belongs to a consumption demand-intensive scenario based on the user's consumption and browsing content. Consumption demand-intensive scenarios include at least one of the following: marriage, car purchase, house purchase, travel, hospitalization, and schooling.

[0040] Then, step 220 is executed: when the user's consumption scenario is a consumption demand-intensive scenario, the user's user information is obtained, and based on the user's user information, the first feature representing the user's merchant preferences, the second feature representing the user's transaction behavior sequence, and the third feature representing the user's consumption scenario are determined.

[0041] Then, step 230 is executed: the first feature, the second feature, and the third feature are input into the user credit limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score;

[0042] Then, step 240 is executed: The user credit limit increase model determines the user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and determines whether to increase the user's credit limit based on the user's comprehensive score;

[0043] Finally, execute step 250: If it is determined that the user's credit limit will be increased, determine the user's increased credit limit based on at least one of the following: consumption potential score, risk level score, and scenario sensitivity score.

[0044] Existing technologies use fewer dimensions to segment users, making it difficult to accurately reflect their true value and needs. In contrast, this application can fully explore user characteristics, making it more accurate in determining whether a user should have their credit limit increased. Furthermore, determining the increased credit limit based on the user's score can meet user needs and reduce financial risks.

[0045] According to one implementation, this application can determine whether a user's consumption scenario belongs to a consumption-intensive scenario based on the user's consumption and browsing content. It should be noted that this application inquires whether the user agrees to provide their information before obtaining the user's consumption and browsing content. The step of obtaining the user's consumption and browsing content is only performed after the user's consent is obtained. The user's consumption and browsing content can be obtained from mobile banking apps, as well as other social media or shopping apps. This application does not limit the specific method of obtaining the user's consumption and browsing content.

[0046] According to one implementation method, intensive consumption demand scenarios can be pre-defined, such as weddings, car purchases, home purchases, travel, hospitalizations, and school enrollment, and may also include other customized scenarios. The difference between intensive consumption demand scenarios and other consumption scenarios is that consumption demand is concentrated in a short period of time, the consumption content is highly related, and the merchants involved are all associated with the same consumption scenario. This application can fully leverage these types of consumption scenarios to determine whether to increase a user's credit limit and the amount of the increase, in order to meet the user's consumption needs within those scenarios.

[0047] According to one implementation method, in determining whether a consumption scenario belongs to a consumption demand-intensive scenario, data collection is first carried out to achieve multi-source consumption and browsing data access, so as to comprehensively and in real time collect user consumption behavior data and browsing behavior data, and provide a data foundation for subsequent judgment.

[0048] Consumer behavior data sources include e-commerce platform transaction records, offline merchant POS machine payment data, bank credit / debit card transaction statements, and transaction details from third-party payment platforms (Alipay, WeChat Pay, etc.). Data fields must include: transaction time, transaction amount, product / service category (goods / service classification code), merchant name and type, payment method, and transaction status (completed / refunded). Browsing behavior data includes browser history, in-app browsing logs (e-commerce apps, news apps, local life apps, etc.), and mini-program browsing history. Data fields must include: browsing time, browsing content title / keywords, browsing page URL / app page identifier, dwell time, click behavior (e.g., viewing details, adding to favorites, adding to cart), and search keywords. Auxiliary data sources include user device information (device ID, operating system), geolocation information (location during consumption / browsing, requiring user authorization), and time-based data (e.g., holidays, semester periods, etc.).

[0049] During feature extraction, specific and general features are extracted for different types of consumption-intensive scenarios (wedding, car purchase, home purchase, travel, hospitalization, schooling) to construct a feature matrix:

[0050] For example, in the context of a wedding, specific characteristics include: frequency / amount of purchases for categories such as wedding photography, wedding services, diamond rings, and wedding dresses; number of times / duration of browsing keywords such as "wedding planning," "wedding venue," and "wedding process"; and the number of interactions with wedding-related merchants in the short term (within 3 months). General characteristics include: total frequency / amount of purchases in the past 3 / 6 months; high-frequency purchase periods; and whether it is a joint purchase by two people (such as multiple user IDs associated with the same order).

[0051] In the context of car purchase, specific characteristics include: purchase records at car dealerships; purchases of auto parts / insurance products; browsing keywords such as "model comparison," "car purchase guide," and "car loan calculation"; and viewing the duration of time spent on car-related apps / pages (a single session exceeding 10 minutes is considered highly relevant). General characteristics include: frequency of large purchases (over 50,000 RMB); whether there has been any browsing behavior related to loan inquiries in the past 6 months; and whether the geographical location frequently appears near car trading markets.

[0052] In the context of home buying, specific characteristics include real estate agency service fees and viewing car service fees; browsing keywords such as "property information," "price trends," "mortgage policies," and "interior design"; the number of times real estate-related pages have been viewed in the past 3 months; general characteristics include: large-scale fund flows in the past year (such as deposit transfers or large-scale transfers); whether or not bank mortgage-related pages have been viewed; and whether the geographical location frequently appears near popular properties / real estate agencies.

[0053] In the context of tourism, specific characteristics include consumption of categories such as air tickets, hotels, attraction tickets, and tour fees; browsing keywords such as "travel guides," "destination guides," and "visa application"; and the frequency of searching for tourism-related content in the past month. General characteristics include whether the consumption time is concentrated around holidays; whether there are cross-city / cross-province consumption plans (such as browsing hotels in other places); and characteristics related to the number of people traveling with the traveler (such as the number of orders being for multiple people).

[0054] In the context of hospitalization, specific characteristics include consumption of categories such as hospital registration fees, hospitalization fees, medicines, and medical devices; browsing keywords such as "disease treatment," "hospitalization process," and "medical insurance reimbursement"; and the frequency of browsing / consuming related to hospitals in the past month. General characteristics include: the total amount of medical-related consumption in the past three months; whether there are consumption / browsing records at the same hospital for several consecutive days; and the frequency of using medical insurance payment methods.

[0055] In the context of schooling, specific characteristics include spending on tuition fees, supplementary teaching materials, school uniforms, and school district housing; browsing keywords such as "admission brochures," "school selection strategies," and "back-to-school preparations"; browsing / spending frequency related to educational institutions (schools, training institutions) in the past 3 months; general characteristics include whether the spending time is concentrated in the back-to-school season (September, February); whether there are usage records of related apps in parent groups; and whether the geographical location is within the school district.

[0056] After feature extraction, feature quantization can be performed: the extracted text features (such as browsing keywords) are converted into numerical features through the bag-of-words model and TF-IDF algorithm, and features such as frequency and amount are normalized (such as Min-Max normalization) to ensure that the magnitude of each feature dimension is consistent.

[0057] Subsequently, scene determination is performed based on the extracted features. Based on the determined feature matrix, an optimized machine learning model can be used to accurately determine scenes with intensive user consumption needs.

[0058] The machine learning model architecture can adopt an end-to-end multi-class machine learning model architecture, directly inputting the processed user behavior feature matrix and outputting the category of the user's consumption demand-intensive scenario (6 categories of intensive / non-intensive scenarios). The model can use gradient boosting tree models (XGBoost, LightGBM) as the base model. These models have good fitting effects on high-dimensional sparse features (such as TF-IDF features of user browsing keywords) and can identify core judgment features through feature importance evaluation, assisting in subsequent feature optimization. Simultaneously, deep learning models (such as multilayer perceptrons, MLPs) are introduced for comparative validation, utilizing their powerful non-linear mapping capabilities to capture complex behavioral patterns. Finally, the model with the better cross-validation score is selected as the final judgment model.

[0059] During training, training data augmentation can be performed: collect large-scale historical user behavior data and complete accurate labeling through "behavioral trajectory tracing + manual review" (for example, if a user subsequently makes a large purchase related to marriage, it is labeled as "wedding scenario"; if there is no intensive scenario purchase for 6 consecutive months, it is labeled as "non-intensive scenario"); construct a training set (70%), a validation set (15%), and a test set (15%), and use stratified sampling to ensure a balanced distribution of samples in each scenario. For scenarios with a small sample size (such as hospitalization scenarios), data augmentation is performed through the SMOTE algorithm to improve the model's ability to judge niche scenarios.

[0060] During model training, multi-class cross-entropy is used as the loss function, and grid search combined with 5-fold cross-validation is used to optimize key model parameters (such as learning rate, tree depth, and number of leaf nodes in LightGBM). L1 regularization is introduced to suppress overfitting, and early stopping strategy is used to avoid performance degradation on the validation set. The model evaluation metric is the F1 score (which balances precision and recall) to ensure balanced judgment performance across various scenarios. The F1 score for core scenarios (such as home purchase and car purchase) must be ≥0.85.

[0061] The final trained consumer demand-intensive scenario recognition model can directly determine that the user belongs to the corresponding consumer demand-intensive scenario if the confidence level of the scenario category output is ≥80%. If the confidence level is <80%, no scenario determination is made, and the user is included in the behavior observation pool. After a period of time, the user's latest consumption and browsing data are collected again, processed, and then input into the model for determination until the confidence level reaches the standard.

[0062] It should be noted that this application inquires whether the user agrees to provide their information before obtaining it. The process of obtaining user information is only carried out after the user's consent. The obtained user information includes basic user information (such as age, occupation, etc.), credit card transaction records (such as transaction time, transaction type, amount, merchant name and merchant type, etc.), repayment records (such as whether repayments were made on time, number of overdue payments, etc.), APP behavioral data (such as login frequency, operation path, etc.), and external data (such as credit reports, social media data, etc.). APP behavioral data specifically refers to data on the user's use of the mobile banking APP. Social media data may specifically include the types of content posted, liked, and followed, in order to infer areas of interest.

[0063] According to one implementation, after obtaining user information, the relevant data of the user information is further cleaned; for example, missing values ​​and outliers are processed to ensure the integrity and accuracy of the data. This application does not limit the steps of data processing.

[0064] According to one implementation, the user information includes credit card transaction records. Determining a first feature representing the user's merchant preferences based on the user's user information includes: determining the merchant type based on the merchant name in the credit card transaction records; and determining the first feature based on the merchant type.

[0065] According to one implementation, credit card transaction records can specifically be implemented as consumption records (or payment records, expenditure records). These records include the transaction details of the party making the transaction via credit card, such as the merchant's name. The merchant name can specifically include the merchant type, such as: dining, shopping, transportation, etc. If the merchant name does not contain the merchant's name itself, relevant information about that merchant can be searched to determine the type of service provided by the merchant, thus establishing the transaction type.

[0066] According to one implementation, if the merchant name includes "restaurant" or "restaurant", transaction records can be extracted to determine that the user's merchant preference characteristics include catering.

[0067] According to one implementation method, when determining the merchant type based on credit card transaction records, user transaction records can be compiled, merchant types can be extracted and standardized, such as supermarkets, restaurants, cinemas, etc.

[0068] According to one implementation, when determining the first feature based on merchant type, the first feature can be determined based on the frequency of occurrence of the merchant type, such as extracting merchant types with a frequency exceeding a preset value as the first feature; the first feature can also be determined in the following ways: converting the merchant names consumed by each user into a merchant type sequence, such as: restaurant-supermarket-cinema-restaurant; using the merchant type sequence as corpus, training a word extraction model; using the trained word extraction model to obtain a vector for each merchant type, such as restaurant vector = [0.1, 0.2, ..., 0.4], where semantically similar merchant type vectors like restaurant and takeout are closer; counting the frequency or amount of user consumption for each merchant type to obtain preference weights, such as supermarket and restaurant consumption ratios of 0.7 and 0.3 respectively; the first feature representing user merchant preference is the preference weight multiplied by the merchant type vector, for example: 0.7 * supermarket vector + 0.3 * restaurant vector.

[0069] This application can accurately determine the type of merchant a user makes a purchase by extracting the merchant name from the transaction record, in order to uncover the user's consumption preferences.

[0070] According to one implementation, determining a second feature representing a user's transaction behavior sequence based on the user's user information includes: determining a target transaction record with transaction continuity based on the transaction time and transaction type in the credit card transaction record; and determining the second feature based on the target transaction record.

[0071] According to one implementation, the acquired user information or user information after data cleaning can be preprocessed to extract statistical features from the target transaction records. The target transaction records include multiple transaction records with similar user behavior patterns, such as all being consumption records or all being repayment records. This application does not limit the similar user behavior patterns possessed by the target transaction records; they can be determined based on transaction time and transaction type, such as the number of consecutive consumption transactions within a period, the maximum amount of a single consumption transaction, the time interval, the repayment cycle, and the fluctuation value of the repayment amount, in order to identify behaviors such as continuous consumption, large-amount consumption, and periodic repayment.

[0072] When determining statistical characteristics of target transaction records, consumption frequency and average consumption amount can be determined based on the transaction records, and on-time repayment rate can be determined based on the repayment records. Secondary characteristics reflect behavioral sequence characteristics, such as consumption trends and repayment pressure indices over the past 7 days determined from credit card transaction records. Behavioral sequence characteristics can reflect a user's continuous consumption characteristics over a period of time, such as consecutive large purchases and periodic repayments. This application identifies continuous target transaction records through statistical analysis of credit card transaction records to determine the user's long-term characteristics.

[0073] According to one implementation, determining a third feature representing a user's consumption scenario based on the user's user information includes: determining a scenario-based tag for each credit card transaction based on credit card transaction records, wherein the scenario-based tag identifies the user's consumption scenario; and determining the third feature based on the scenario-based tags corresponding to the credit card transactions included in the credit card transaction records.

[0074] According to one implementation method, scenario-based tags can be pre-constructed, including business travel scenarios, promotional scenarios, and daily consumption scenarios. For a user's credit card transaction records, corresponding scenario-based tags can be determined. For example, a flight ticket purchase transaction can be identified as a business travel scenario; a transaction near Double Eleven or Double Twelve can be identified as a promotional scenario; if it cannot be determined whether it is a business travel consumption tag or a promotional scenario tag, it can be identified as a daily consumption tag. After determining the scenario-based features of each user's credit card transaction record, a scenario-based feature set or scenario-based feature sequence is determined as a third feature based on the user's multiple credit card transaction records. The scenario-based feature set refers to which scenario-based tags are included in the user's multiple credit card transaction records; the scenario-based feature sequence is the scenario-based sequence feature determined after sorting the user's multiple credit card transaction records. This application identifies the consumption scenario of each credit card transaction to explore the user's sensitivity to consumption scenarios.

[0075] According to one implementation, obtaining user information includes: determining the size of a time window based on the user's frequency of operation on the online banking application or the user's transaction frequency, wherein the higher the user's frequency of operation on the online banking application or the user's transaction frequency, the smaller the time window; determining the time range for obtaining user information based on the size of the time window, and obtaining user information based on the time range.

[0076] According to one implementation, when determining features such as the first feature, the second feature, and the third feature, a dynamic time window for acquiring user information can be set to extract the volatility of consumption during feature determination. Consumption volatility includes changes in the number of transactions, amount, etc., within the specified window. Consumption volatility can be used to distinguish between short-term large-amount consumption and long-term stable small-amount consumption behavior, thereby setting different recommendation strategies.

[0077] A dynamic time window allows for flexible setting of the time window size, such as 5 minutes, 1 hour, 1 day, 7 days, or 30 days. The time window size can be determined by considering parameters such as user activity, model performance, and user lifecycle stage. For example, if a user has recently engaged in frequent consumption or login activities, the window can be shortened to 7 days to capture short-term behavioral trends. If user behavior is relatively stable, the window can be expanded to 30 days or more to avoid feature fluctuations due to insufficient data. If the model's classification results fluctuate frequently, the window needs to be expanded to smooth the features. If the model's response delay to new behaviors is too high, the window can be reduced to improve real-time performance. In one implementation, a small window can be used for rapid learning of new users, while a large window can be used for stable classification of existing users. This application flexibly adjusts the time window size based on the user's operation frequency, ensuring sufficient data for user feature mining and increasing the frequency of determining whether to increase user credit limits, thus enabling more timely credit limit increases.

[0078] According to one implementation, within a time window, data from different times can be assigned different weights; for example, more recent data can be assigned a higher weight, while data from more distant times can be assigned a lower weight. This application does not limit the specific method for assigning weights to data from different times.

[0079] According to one implementation, inputting a first feature, a second feature, and a third feature into a user credit limit increase model to determine a user's consumption potential score, risk level score, and scenario sensitivity score includes: determining a user's consumption potential score for different merchant types based on the first and second features; determining a user's risk level score based on the second feature and the user's credit card repayment history; and determining a user's scenario sensitivity score based on the first and third features, wherein the scenario sensitivity score describes the user's willingness to consume in different consumption scenarios.

[0080] According to one implementation, this application can pre-construct training samples, labeling consumption potential scores, risk level scores, and scenario sensitivity scores to obtain a user credit limit increase model capable of performing multiple tasks. The user credit limit increase model receives multimodal features, including first features, second features, and third features, etc.; and determines the user's scores across multiple dimensions based on the multimodal features.

[0081] When applying the user credit limit increase model, the first feature represents the user's merchant preferences, and the second feature represents the user's transaction behavior sequence. The user credit limit increase model can learn the correlation between the first and second features, such as a user's multiple purchases at restaurants or clothing stores, to determine the user's spending potential score for different merchant types based on the first and second features. The spending potential score describes the likelihood of the user making future purchases at different merchants. The spending potential score can also describe the probability of the user's future spending, for example, by weighting and summing the probabilities of future spending at different merchants.

[0082] The user credit limit increase model determines a user's risk level score based on a second characteristic and the user's credit card repayment history. The second characteristic includes the user's monthly spending, and the debt ratio can be calculated based on this characteristic and the user's credit card repayment history, all used to calculate the risk level score. The risk level score describes the risk of the user defaulting or being overdue on repayments after a credit limit increase.

[0083] The user credit limit increase model determines a user's scenario sensitivity score based on the first and third features. The third feature describes the user's consumption scenario; the model learns the correlation between the first and third features to determine whether a user is willing to consume in different scenarios. For example, in business travel scenarios, users prefer to consume at merchants in the catering, shopping, and transportation sectors; or in promotional scenarios (such as Singles' Day or Double Twelve), users prefer to consume at merchants in the clothing sector. When determining a user's scenario sensitivity score based on the first and third features, factors such as the percentage of promotional orders from different merchants relative to the total annual orders, the percentage of orders using coupons, the click-through rate of promotional pushes, and the frequency of price comparison behavior can be considered. Higher values ​​for these indicators indicate that the user is more sensitive to promotional scenarios, resulting in a higher scenario sensitivity score.

[0084] This application utilizes cross-learning of multiple features to more comprehensively uncover the correlations between different features, more accurately determine relevant scores, and whether to increase the user's credit limit.

[0085] According to one implementation method, after determining a user's consumption potential score, risk level score, and scenario sensitivity score, the user credit limit increase model determines a user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and determines whether to increase the user's credit limit based on the user's comprehensive score.

[0086] According to one implementation, the user credit limit increase model can set different weights for consumption potential score, risk level score and scenario sensitivity score, and then weight the different scores according to the different weights set for each score to obtain the user's comprehensive score.

[0087] According to one implementation, the user credit limit increase model can assign weights of 0.45, 0.35, and 0.2 to the consumption potential score, risk level score, and scenario sensitivity score, respectively. These weights can be implemented as initial weights and flexibly adjusted in subsequent processes based on the model's determination of whether to increase the credit limit, in order to improve model performance.

[0088] According to one implementation method, after obtaining a user's comprehensive score, it can be determined whether to increase the user's credit limit based on the relationship between the user's comprehensive score and a preset threshold. If the user's comprehensive score is greater than the preset threshold, it indicates that the user has good credit and a high willingness to spend, and the user's credit limit can be increased; conversely, if the user's comprehensive score is less than the preset threshold, the user's credit limit will not be increased in this instance.

[0089] According to one implementation method, when it is determined that a user's credit limit will be increased, the increased credit limit is determined based on at least one of a consumption potential score, a risk level score, and a scenario sensitivity score.

[0090] According to one implementation method, the credit limit after a user's credit limit increase can be determined based on the consumption potential score and / or risk level score; the credit limit after a user's credit limit increase can also be determined based on the scenario sensitivity score, for example, in a specific consumption scenario (business travel scenario or promotional scenario), a temporary credit limit can be set for the user, as well as the size of the temporary credit limit.

[0091] According to one implementation method, the user's credit limit in a specific consumption scenario can also be determined based on the consumption potential score and the scenario sensitivity score; or the risk level score and the scenario sensitivity score; or the credit limit can be determined by comprehensively considering the consumption potential score, the risk level score, and the scenario sensitivity score; or the user's credit limit after the credit limit increase can be determined directly based on the user's comprehensive score.

[0092] According to one implementation, this application can also execute this method on users during a specific time period to determine whether to increase the user's credit limit; for example, before Double Eleven, Double Twelve, holidays, etc., this method can be executed to determine whether to increase the user's credit limit to meet the user's consumption needs in promotional or holiday scenarios.

[0093] According to one implementation, this application can also determine whether to increase the credit limit based on the frequency of a user's credit card spending. For example, if a user's spending frequency reaches 10 times in the past 7 days or 3 days, the application identifies the scenarios in which the user might have made the purchases based on the content of their spending, and then executes this method to determine whether to increase the user's credit limit. For example, if a user's spending frequency reaches 10 times in the past 7 days or 3 days, and the spending includes transportation, then the user's spending scenario may be a business travel scenario. In this case, the application executes this method to determine whether to increase the user's credit limit to meet the user's spending needs in business travel scenarios.

[0094] According to one implementation, this application can also determine whether to increase a user's credit limit based on the user's usage and browsing data across different applications. For example, if a user has recently browsed wedding products or car products extensively, or has recently commented extensively on content related to infants and toddlers, it can be determined that the user may have wedding plans, car purchase plans, or childcare plans; subsequently, this method is executed to determine whether to increase the user's credit limit to meet the user's consumption plans. It should be noted that the different applications obtained in this application can specifically be applications developed by banks or applications that have established data exchange agreements with banks. Furthermore, regardless of the application whose data is obtained, the user's consent is obtained before acquiring this data. When determining a user's consumption plans, this application can utilize natural language processing technology to perform sentiment analysis on user feedback text in social media data to extract the user's sentiment orientation. User feedback text includes user evaluations on shopping, social media, and other software; user sentiment orientation includes the user's consumption intention for a product, such as whether they expect to buy or not.

[0095] According to one implementation method, a fourth feature can also be input to determine whether to increase the user's credit limit. The fourth feature includes a credit profile determined for the user. The credit profile includes multiple credit characteristics determined for the user, such as: integrating data from the central bank's credit system, Baihang Credit System, and internal bank data to generate a comprehensive credit score; calculating credit change trends (such as the rate of change in the number of credit inquiries in the past six months); and developing a default risk index (combining historical overdue duration, amount, and frequency), etc.

[0096] According to one implementation, this application also performs model verification and iteration, including setting evaluation indicators such as silhouette coefficient, inter-cluster variability and business indicators to evaluate the model. The business indicators include: variance of user response rate after stratification; and setting a continuous optimization mechanism: weekly model drift detection, automatic retraining triggered by abnormal clusters, and adjustment of feature weights based on expert feedback.

[0097] According to one implementation, this application also sets up a real-time credit limit adjustment recommendation engine to initiate the execution of this method, improving the timeliness of determining whether to increase the user's credit limit. This engine integrates online learning, reinforcement decision-making, and causal reasoning technologies to achieve millisecond-level credit limit adjustment strategy generation. Its core technical architecture is as follows: Real-time data access and feature calculation, including: setting up a data pipeline: using a distributed publish-subscribe messaging system to build a high-throughput message queue, accessing real-time transaction data (consumption timestamp, merchant type, amount), repayment behavior (overdue status, minimum repayment ratio), device information (login location, terminal type), etc.; integrating an open-source stream processing framework for real-time data cleaning, handling missing values ​​and abnormal transactions (such as high-frequency small-amount test transactions); setting up real-time feature engineering, including: sliding window calculation: based on the behavioral data of the last 5 minutes / 1 hour / 1 day, generating dynamic features such as consumption frequency and repayment pressure index (current debt / monthly income); real-time graph calculation: using a graph database to build a user relationship graph, identifying consumption linkage features within a friend's circle; real-time storage: writing features into a database, supporting microsecond-level queries.

[0098] According to one implementation, an online model inference system is also set up. This includes: lightweight model deployment: deploying a response probability prediction model using a lightweight machine learning framework, compressing the model size to below 5MB; real-time graph model inference based on a distributed graph database to identify sudden changes in user consumption patterns (such as a surge in business travel spending); specifically, a graph database can be used, connecting to real-time transaction data, such as consumption time, type, and amount, and user-related data, such as social friends and shared orders, to construct a user-family / friends network graph. When a user makes a transaction, the behavior is converted into network nodes and links in real time. The graph is analyzed using graph algorithms, clustered according to relationship density, and family / friends communities with similar relationships are identified. Linkage indicators for consumption time, category, and scenario are calculated for these family / friends communities to determine whether to implement this method to increase the user's credit limit.

[0099] This application also enables hot updates of the model, including: using A / B testing services to achieve seamless model version switching; and a triggered update mechanism: when a user behavior distribution offset is detected to exceed a threshold (such as KS value > 0.2), the latest model is automatically loaded.

[0100] According to one implementation, this application also includes a dynamic strategy generator, comprising a reinforcement learning decision module, which uses a deep reinforcement learning framework to construct a multi-armed slot machine model; it also defines the action space: credit limit adjustment range (10%~50% step size), marketing channels (SMS / APP push), time window (morning / evening); and a reward function design: taking into account factors such as credit limit adjustment response rate (real-time feedback), user retention rate (7-day window), and risk coefficient (30-day delinquency rate after loan) to set the reward function in order to optimize the model.

[0101] According to one implementation, a causal effect calculator is also provided, for example, to calculate the propensity score matching similar user groups in real time, and to use a causal forest model to predict the long-term effects of different credit adjustment strategies (such as the consumption increase rate after 6 months) in order to adjust the model.

[0102] According to one implementation, when adjusting the model, a real-time feedback closed loop is also set up, including response behavior capture: for example, real-time collection of user behavior data such as clicking on the credit limit adjustment notification and actual card usage through the tracking SDK; establishing a behavior-result mapping table to record the response delay distribution 30 minutes / 1 hour / 1 day after the credit limit adjustment operation; the strategy dynamic optimization also sets up a real-time data stream warehouse based on Delta Lake to achieve minute-level evaluation of the strategy effect, and uses evolutionary algorithms (such as genetic algorithms) to dynamically adjust the strategy parameters and generate Pareto optimal solution sets to assist in the adjustment of model parameters.

[0103] Based on the above-mentioned method for determining a user's credit card limit, this application also provides a device for determining a user's credit card limit. The following will combine... Figure 3 The device is described in detail.

[0104] Figure 3 A schematic block diagram of a user credit card limit determination device according to an embodiment of this application is shown.

[0105] like Figure 3 As shown, the user credit card limit determination device 300 of this embodiment includes a consumption demand intensive scene recognition module 310, an initial feature extraction module 320, a score calculation module 330, a comprehensive score determination module 340, and a credit card limit adjustment module 350.

[0106] The consumer demand-intensive scenario identification module is used to determine whether a user's consumption scenario belongs to a consumer demand-intensive scenario based on the user's consumption and browsing content. The consumer demand-intensive scenario includes at least one of the following: getting married, buying a car, buying a house, traveling, hospitalization, and going to school.

[0107] The initial feature extraction module 320 is used to obtain user information when the user's consumption scenario belongs to the consumption demand-intensive scenario, and to determine the first feature representing the user's merchant preference, the second feature representing the user's transaction behavior sequence and the third feature representing the user's consumption scenario based on the user information.

[0108] User information includes credit card transaction records. The initial feature extraction module 320 includes a first feature determination module, which is used to determine a first feature representing the user's merchant preferences based on the user's user information, including: determining the merchant type based on the merchant name in the credit card transaction records; and determining the first feature based on the merchant type.

[0109] The initial feature extraction module 320 also includes a second feature determination module, which is used to determine a second feature representing the user's transaction behavior sequence based on the user's user information, including: determining a target transaction record with transaction continuity based on the transaction time and transaction type in the credit card transaction record; and determining the second feature based on the target transaction record.

[0110] The initial feature extraction module 320 also includes a third feature determination module, which is used to determine the third feature representing the user's consumption scenario based on the user's user information, including: determining the scenario-based label of each credit card transaction based on the credit card transaction record, the scenario-based label identifying the user's consumption scenario; and determining the third feature based on the scenario-based labels corresponding to the credit card transactions included in the credit card transaction record.

[0111] The initial feature extraction module 320 also includes a user information acquisition module, which is used to determine the size of the time window based on the user's operation frequency or transaction frequency in the online banking application. The higher the user's operation frequency or transaction frequency in the online banking application, the smaller the time window. The time range for acquiring user information is determined based on the size of the time window, and user information is acquired based on the time range.

[0112] The scoring calculation module 330 is used to input the first feature, the second feature and the third feature into the user credit limit increase model to determine the user's consumption potential score, risk level score and scenario sensitivity score;

[0113] The scoring calculation module 330 is also used to determine the user's consumption potential score for different merchant types based on the first feature and the second feature; to determine the user's risk level score based on the second feature and the user's credit card repayment record; and to determine the user's scenario sensitivity score based on the first feature and the third feature. The scenario sensitivity score describes the user's willingness to consume in different consumption scenarios. The comprehensive score determination module 340 is used to call the user credit limit increase model to determine the user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and to determine whether to increase the user's credit limit based on the user's comprehensive score.

[0114] The credit card limit adjustment module 350 is used to determine the user's increased credit limit based on at least one of the following: spending potential score, risk level score, and scenario sensitivity score, when it is determined that the user's credit limit will be increased.

[0115] According to embodiments of this application, any multiple modules among the consumer demand-intensive scene recognition module 310, initial feature extraction module 320, score calculation module 330, comprehensive score determination module 340, and credit card limit adjustment module 350 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some functions of one or more of these modules can be combined with at least some functions of other modules and implemented in one module. According to embodiments of this application, at least one of the consumer demand-intensive scene recognition module 310, initial feature extraction module 320, score calculation module 330, comprehensive score determination module 340, and credit card limit adjustment module 350 can be at least partially implemented as hardware circuits, such as field-programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems-on-a-chip, systems-on-a-substrate, systems-on-package, application-specific integrated circuits (ASICs), or any other reasonable means of integrating or packaging circuits, or implemented in software, hardware, or firmware, or in any appropriate combination of any of these three implementation methods. Alternatively, at least one of the consumer demand-intensive scene recognition module 310, initial feature extraction module 320, score calculation module 330, comprehensive score determination module 340, and credit card limit adjustment module 350 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0116] Figure 4 A block diagram schematically illustrates an electronic device suitable for implementing a user credit card limit determination method according to an embodiment of this application.

[0117] like Figure 4 As shown, an electronic device 400 according to an embodiment of this application includes a processor 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage portion 408 into a random access memory (RAM) 403. The processor 401 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 401 may also include onboard memory for caching purposes. The processor 401 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0118] RAM 403 stores various programs and data required for the operation of electronic device 400. Processor 401, ROM 402, and RAM 403 are interconnected via bus 404. Processor 401 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 402 and / or RAM 403. It should be noted that the programs may also be stored in one or more memories other than ROM 402 and RAM 403. Processor 401 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0119] According to embodiments of this application, the electronic device 400 may further include an input / output (I / O) interface 405, which is also connected to a bus 404. The electronic device 400 may also include one or more of the following components connected to the input / output (I / O) interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output (I / O) interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 410 as needed so that computer programs read from it can be installed into the storage section 408 as needed.

[0120] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0121] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 402 and / or RAM 403 and / or one or more memories other than ROM 402 and RAM 403 described above.

[0122] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the user credit card limit determination method provided in the embodiments of this application.

[0123] When the computer program is executed by the processor 401, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0124] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via communication section 409, and / or installed from removable medium 411. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0125] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by processor 401, it performs the functions defined in the system of this application embodiment. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0126] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0127] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0128] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

[0129] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A method for determining a user's credit card limit, characterized in that, The method includes: The user's consumption scenario is determined based on their consumption and browsing content. The consumption scenario includes at least one of the following: marriage, car purchase, house purchase, travel, hospitalization, and schooling. When the user's consumption scenario is a consumption demand-intensive scenario, the user's user information is obtained, and based on the user's user information, a first feature representing the user's merchant preferences, a second feature representing the user's transaction behavior sequence, and a third feature representing the user's consumption scenario are determined. Input the first feature, the second feature, and the third feature into the user credit limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score; The user credit limit increase model determines a user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and determines whether to increase the user's credit limit based on the comprehensive user score. If it is determined that the user's credit limit will be increased, the increased credit limit will be determined based on at least one of the consumption potential score, risk level score, and scenario sensitivity score.

2. The method according to claim 1, wherein, The user information includes credit card transaction records. Determining a first feature characterizing the user's merchant preferences based on the user information includes: The merchant type is determined based on the merchant name in the credit card transaction record; The first feature is determined based on the merchant type.

3. The method according to claim 2, wherein, Determining a second feature characterizing the user's transaction behavior sequence based on the user's user information includes: Target transaction records with transaction continuity are determined based on the transaction time and transaction type in the credit card transaction records; The second feature is determined based on the target transaction record.

4. The method according to any one of claims 1-3, wherein, The third feature representing the user's consumption scenario, determined based on the user's information, includes: Each credit card transaction is tagged with a contextual label based on the credit card transaction records, and the contextual label identifies the user's consumption scenario. The third feature is determined based on the contextualized tags corresponding to the credit card transactions included in the credit card transaction records.

5. The method according to any one of claims 1-3, wherein, The user information obtained includes: The size of the time window is determined based on the frequency of user operations or transactions on the online banking application. The higher the frequency of user operations or transactions on the online banking application, the smaller the time window. The time range for obtaining user information is determined based on the size of the time window, and the user information is obtained based on the time range.

6. The method according to any one of claims 1-3, wherein, The step of inputting the first feature, the second feature, and the third feature into the user credit limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score includes: Based on the first feature and the second feature, determine the user's consumption potential score for different merchant types; The user's risk level score is determined based on the second feature and the user's credit card repayment history; The user's scenario sensitivity score is determined based on the first feature and the third feature, and the scenario sensitivity score describes the user's willingness to consume in different consumption scenarios.

7. A device for determining a user's credit card limit, characterized in that, The device includes: The consumption demand intensive scenario identification module is used to determine whether a user's consumption scenario belongs to a consumption demand intensive scenario based on the user's consumption and browsing content. The consumption demand intensive scenario includes at least one of the following: marriage, car purchase, house purchase, travel, hospitalization, and schooling. The initial feature extraction module, when the user's consumption scenario belongs to a consumption demand-intensive scenario, obtains the user's user information, and determines a first feature representing the user's merchant preferences, a second feature representing the user's transaction behavior sequence, and a third feature representing the user's consumption scenario based on the user's user information. The scoring calculation module is used to input the first feature, the second feature, and the third feature into the user credit limit increase model to determine the user's consumption potential score, risk level score, and scenario sensitivity score. The comprehensive score determination module is used to call the user credit limit increase model to determine the user's comprehensive score based on the consumption potential score, risk level score, and scenario sensitivity score, and to determine whether to increase the user's credit limit based on the user's comprehensive score; The credit card limit adjustment module, when it is determined that the user's limit will be increased, determines the increased limit based on at least one of the following: the spending potential score, the risk level score, and the scenario sensitivity score.

8. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1-6.

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

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1-6.

Images