Account quota dynamic adjustment method and device, computer equipment and storage medium

Abstract

The invention discloses an account quota dynamic adjustment method and device, computer equipment and a storage medium. The method comprises the steps of obtaining account related information, and performing cleaning and filtering to obtain a preprocessing result; based on the preprocessing result, the account is classified as active or inactive according to the account activity condition, so that the account type is obtained; based on the account type, adopting a corresponding risk rating model to perform risk rating, and identifying a potential risk to obtain an account risk; calculating customer value scores, and dividing value grades to obtain customer value grades; automatically adjusting the account transaction limit to obtain an adjustment result; and synchronizing to a related service system. By implementing the method provided by the invention, more intelligent and dynamic account quota adjustment with high adaptability can be realized, so that the problems that account quota adjustment in the prior art lacks real-time performance and is not flexible enough, normal customers and potential risk customers cannot be effectively distinguished, misjudgment is caused to high-value customers and the like are solved.

Description

Technical Field

[0001] This invention relates to data processing methods, and more specifically to methods, apparatus, computer equipment, and storage media for dynamically adjusting account limits. Background Technology

[0002] With the development of information technology and the widespread adoption of mobile internet, financial services, especially electronic payment services, have become increasingly convenient. However, this has also led to increased security risks, particularly in account management and transaction limit control. Traditional methods for adjusting account limits are no longer sufficient to meet the needs of modern financial operations, especially in combating illegal activities such as telecommunications fraud and gambling.

[0003] The existing "5G-based account limit adjustment scheme" allows customers to receive limit adjustment messages from the bank's server via their terminal devices for identity verification and limit adjustment. While this method simplifies the customer's process, its main limitation is its inability to respond in real time to changes in customer assets and credit status, and its inability to provide personalized limit management based on different customer groups and scenarios. Furthermore, this scheme is inefficient when processing batches of customer limit adjustment requests and struggles to meet the demands of a rapidly changing risk environment. The "electronic channel transaction limit control system" is primarily used for real-time verification and control of customer transaction limits in electronic channel transactions. Although it effectively limits transactions exceeding preset limits, it lacks flexibility when dealing with different customer groups and applicable scenarios, failing to achieve precise risk control. Moreover, this system relies on fixed cumulative limit settings, which can easily lead to unnecessary transaction failures or restrictions for customers with highly volatile transaction behaviors, impacting customer experience. The "customer account limit adjustment request method" utilizes a federated learning model to predict and adjust customer account limits, aiming to improve the personalization and efficiency of account limit adjustments. However, it also faces the problem of untimely data updates, making it difficult to reflect the customer's latest financial situation and risk status in real time. Moreover, the lack of specific analysis targeting particular risk characteristics, such as credit card farming behavior, limits the effectiveness of this method in identifying and preventing fraudulent activities.

[0004] Therefore, it is necessary to design a new method to achieve more intelligent, dynamic and highly adaptive account limit adjustments in order to solve the problems of existing technologies in account limit adjustments, such as lack of real-time performance, insufficient flexibility, inability to effectively distinguish between normal customers and potential risk customers, and misjudgment of high-value customers. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, apparatus, computer equipment and storage medium for dynamic adjustment of account limits.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for dynamically adjusting account limits, comprising: Obtain account-related information and perform cleaning and filtering to obtain preprocessed results; Based on the preprocessing results, the accounts are classified as active or inactive according to their activity levels to obtain the account type; Based on the account type, the corresponding risk rating model is used to assess the risk level and identify potential risks in order to obtain the account risk. The preprocessing results are analyzed to determine the customer's assets and transaction frequency indicators, calculate the customer value score, and classify the customer into value levels to obtain the customer value level. The account transaction limit is automatically adjusted based on the account risk and the customer value level to obtain the adjustment result; The adjustment results will be synchronized to the relevant business systems.

[0007] The further technical solution is as follows: the acquisition of account-related information and its cleaning and filtering to obtain preprocessing results include: Obtain account-related information required for account risk and customer value assessment, cleanse, transform and process the account-related information, and filter out transactions not initiated by the customer to obtain preprocessed results.

[0008] The further technical solution is as follows: The account's activity level is categorized based on factors such as the cumulative transaction volume and frequency on a given day. For example, an account with transactions on a given day, a cumulative transaction amount of 1000 yuan or more, at least 3 outgoing transactions, and at least 1 incoming transaction is considered an active account; conversely, accounts that do not meet the criteria for an active account are considered inactive accounts. The criteria for determining activity level can be adjusted according to actual business needs.

[0009] The further technical solution is as follows: Based on the account type, a corresponding risk rating model is used to assess the risk level and identify potential risks to obtain account risk, including: The preprocessing results are then processed using risk indicators to obtain account classifications; Determine the appropriate risk assessment model based on the account type; Based on the account classification, the accounts are input into the risk assessment model to identify potential risks and obtain the risk assessment results for this period. By comparing the current assessment results with the historical risk rating results, the risk classification and grading of the account are updated according to the established logic to obtain the account risk.

[0010] The further technical solution is as follows: the setting strategy includes a list strategy, an abnormal probing behavior strategy, an abnormal operation behavior strategy, an abnormal transaction behavior strategy, due diligence risk identification strategy, an abnormal change behavior strategy, behavior correlation mining strategy, attention strategy, and trust strategy.

[0011] The list strategy is to quickly identify and manage risks by matching and updating whitelists or risk lists, ensuring compliance and supporting flexible configuration to adapt to different business needs; The abnormal probing behavior strategy uses rules, statistical methods, and machine learning techniques to analyze customer behavior and transaction records in order to discover potential probing risk behaviors and assess the risk situation. Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal trading behavior strategies are based on modeling normal trading behavior and use methods such as machine learning and statistical analysis to identify trading behaviors that deviate from normal patterns and assess the risk situation. The due diligence risk identification strategy combines expert knowledge and data analysis tools, and collects information through interviews, questionnaires and other methods to identify potential risks and assess the risk situation. The abnormal behavior change strategy involves monitoring behavioral data and analyzing historical data, and establishing a baseline of normal behavior to identify abnormal behaviors that deviate from the normal pattern and assess the risk situation. Behavioral association mining strategy uses data mining techniques to analyze data from multiple sources, uncover association rules between different behaviors to identify potential risk patterns, and assess risk situations; The focus strategy involves continuous monitoring of risks that cannot be completely avoided or effectively controlled, so that swift action can be taken when the risks first emerge to reduce their impact on the business. Trustworthy strategies prioritize identifying high-quality accounts based on trustworthy characteristics, improving the accuracy and efficiency of decision-making, avoiding strict spending limits on these high-quality accounts, and can be dynamically adjusted according to actual circumstances.

[0012] Its further technical solution is as follows: Analyze customer assets and transaction frequency indicators, calculate customer value scores, and classify them into value levels to obtain customer value levels, including: We collect relevant information from multiple dimensions, including customer assets, credit, and account activity. The relevant information is processed, quantified and stratified according to the characteristics that measure customer value, to obtain indicators; Based on the degree of impact of the indicators on customer value, assign weights to each indicator; Calculate the customer's overall value score using a weighted summation method; Based on the comprehensive value score, customers are divided into three or more value levels: high, medium, and low, to obtain customer value levels.

[0013] The further technical solution is as follows: based on the account risk and the customer value level, automatically match the corresponding dynamic limit adjustment strategy to automatically adjust the account's transaction limit, including: Based on the account risk and the customer value level, different fixed or dynamic limit reduction schemes are determined, for example: implementing limit control risk for high-risk, low-value accounts; Based on the account risk and the customer value level, a tiered credit limit increase rule is designed to provide corresponding credit limit increases for accounts of different levels. For example, increasing the credit limit for low-risk, high-value accounts can improve the customer experience. When account risk or customer value level changes in real time, the credit limit adjustment process is automatically triggered. Based on the defined matrix credit limit adjustment plan and strategy, combined with the current account credit limit, the operation of reducing or increasing the credit limit is automatically generated and executed to obtain the adjustment result.

[0014] The present invention also provides a device for dynamically adjusting account limits, comprising: The acquisition unit is used to acquire account-related information and perform cleaning and filtering to obtain preprocessed results; A classification unit is used to classify accounts as active or inactive based on the preprocessing results and their activity status to obtain the account type; The risk assessment unit is used to assess the risk level based on the account type using the corresponding risk rating model, identify potential risks, and obtain the account risk. The rating determination unit is used to analyze the customer's assets and transaction frequency indicators based on the preprocessing results, calculate the customer value score, and classify the value levels to obtain the customer value level. An adjustment unit is used to automatically adjust the account transaction limit based on the account risk and the customer value level to obtain the adjustment result; The synchronization unit is used to synchronize the adjustment results to the relevant business systems.

[0015] The present invention also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the above-described method.

[0016] The present invention also provides a storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0017] The advantages of this invention compared to existing technologies are as follows: This invention acquires and preprocesses account-related information, classifies accounts into active or inactive types based on their activity, applies corresponding risk rating models to different types to identify potential risks, and analyzes indicators such as customer assets and transaction frequency to calculate customer value scores. Finally, it automatically adjusts account transaction limits based on account risk and customer value level, and synchronizes the adjustment results to relevant business systems, thus achieving a more intelligent, dynamic, and highly adaptable account limit adjustment mechanism. This method solves the problems of existing technologies, such as lack of real-time performance, insufficient flexibility, difficulty in effectively distinguishing between normal and potentially risky customers, and misjudgment of high-value customers. It achieves the goal of meeting the needs of different customers while ensuring security, improving the accuracy of risk management and service quality. Through this comprehensive assessment and automated adjustment approach, it not only enhances the ability to prevent fraudulent and other risky behaviors but also ensures a better service experience for high-value customers.

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram illustrating an application scenario of the dynamic account limit adjustment method provided in this embodiment of the invention. Figure 2 A flowchart illustrating the method for dynamically adjusting account limits provided in an embodiment of the present invention; Figure 3 A schematic diagram of a sub-process of the dynamic adjustment method for account limits provided in an embodiment of the present invention; Figure 4 A schematic diagram of a sub-process of the dynamic adjustment method for account limits provided in an embodiment of the present invention; Figure 5 A schematic diagram of a sub-process of the dynamic adjustment method for account limits provided in an embodiment of the present invention; Figure 6 A schematic block diagram of an account limit dynamic adjustment device provided in an embodiment of the present invention; Figure 7 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0023] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0024] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0025] Please see Figure 1 and Figure 2 , Figure 1 This is a schematic diagram illustrating an application scenario of the dynamic account limit adjustment method provided in this embodiment of the invention. Figure 2 This is a schematic flowchart illustrating the dynamic account limit adjustment method provided in this embodiment of the invention. The method is applied in a server. The server interacts with the terminal, integrating, cleaning, and classifying account-related information. It distinguishes between active and inactive accounts based on account activity, employs a targeted risk rating model to assess potential risks, and analyzes customer assets and transaction frequency to calculate a customer value score. The transaction limit is then automatically adjusted based on account risk and customer value level. This method utilizes various strategies (such as list-based strategies and abnormal behavior identification strategies) for detailed risk assessment and classifies customers into different levels based on the comprehensive value score. This achieves precise control over high-risk, low-value accounts and improves the customer experience for low-risk, high-value accounts. The entire process emphasizes real-time performance, flexibility, and adaptability, addressing the shortcomings of existing technologies in account limit adjustment, such as lack of real-time updates, insufficient flexibility, and difficulty in effectively distinguishing between normal and potentially risky customers. It particularly avoids misjudging high-value customers, thus providing a more intelligent, dynamic, and highly adaptable account management solution.

[0026] Figure 2 This is a flowchart illustrating the method for dynamically adjusting account limits provided in an embodiment of the present invention. Figure 2 As shown, the method includes the following steps S110 to S160.

[0027] S110. Obtain account-related information and perform cleaning and filtering to obtain preprocessed results.

[0028] In this embodiment, preprocessing results refer to high-quality datasets obtained after a series of data processing steps that can be directly used for subsequent analysis and model calculations. Specifically, these preprocessing results include the following characteristics: Accuracy: Errors and inconsistencies in all raw data have been corrected or removed, ensuring the authenticity and reliability of the information.

[0029] Relevance: Only data directly related to account risk assessment and customer value evaluation is retained, while irrelevant or redundant information, such as transactions not initiated by the customer, is removed.

[0030] Completeness: Missing data was appropriately filled in or marked to ensure that each data record is complete for subsequent analysis.

[0031] Standardization: Data from different sources is converted into the same format and units to facilitate comparison and integration.

[0032] Derivative variables: New variables or indicators are generated based on the original data, such as cumulative transaction amount and cumulative number of outgoing transactions. These derived variables help to more comprehensively describe the behavioral characteristics and risk status of the account.

[0033] Cleaning and filtering: Filter out transactions that may interfere with the calculation of risk indicators through specific rules (such as transaction type, channel, and product type) to ensure that the remaining data accurately reflects the actual behavior and intentions of customers.

[0034] In summary, the preprocessing results form the foundation of the entire intelligent dynamic credit limit adjustment solution, providing reliable data support for subsequent modules, including but not limited to account classification, risk rating, and customer value assessment. This carefully prepared dataset significantly improves the accuracy and efficiency of subsequent analysis, thereby enhancing the accuracy and effectiveness of the entire system's decision-making.

[0035] Specifically, the process involves obtaining account-related information required for account risk and customer value assessment, cleaning, transforming, and processing the account-related information, and filtering out transactions not initiated by the customer to obtain preprocessed results.

[0036] Specifically, this process first involves collecting various data sources related to the account, including but not limited to basic account and customer information (such as age, occupation, etc.), customer contribution and credit information, account fund transaction details, account behavior data (such as login frequency, operation type, etc.), and external data (such as credit score). This data is collected from multiple platforms and needs to undergo rigorous screening and processing to adapt to specific business needs and indicator requirements.

[0037] The next step is to clean and transform the collected data. The cleaning process primarily aims to remove or correct errors and inconsistencies in the data, such as handling missing values, duplicate records, and formatting errors. Furthermore, the data will be transformed and processed according to actual needs, such as mapping the raw data to a more suitable format for analysis, performing necessary calculations to generate new derived variables, or standardizing data from different sources according to a unified standard.

[0038] Specifically, at this stage, any transactions not initiated by the customer will be identified and filtered out to avoid unnecessary interference with risk assessment and value scoring. These transactions may include reversal transactions, bank fee transactions, interest settlement transactions, entrusted payment transactions, and refund transactions, which typically do not reflect the customer's true behavior or intentions. Determining which transactions should be excluded based on three dimensions—transaction type, channel, and product type—ensures that the remaining data is cleaner and more representative.

[0039] Finally, after the aforementioned series of data processing steps, preprocessed results are obtained. These results not only include high-quality data that has been cleaned, transformed, and processed, but also directly support the functional implementation of subsequent modules, such as account classification, risk rating, and customer value assessment, which is crucial for improving the accuracy of the entire system's decision-making. Therefore, the successful execution of this step is key to ensuring the effective operation of each subsequent stage.

[0040] Specifically, firstly, technicians clean, transform, and process the data to be input into the model. This step is crucial because it directly affects the accuracy and reliability of subsequent analysis results. Specifically: Data cleaning: removing errors, duplicates, and inconsistencies from the original dataset.

[0041] Data conversion: unifying data from different sources into the same format and units to facilitate comparison and integration.

[0042] Data processing: Generating new variables or indicators based on existing data to help more comprehensively describe the characteristics of account behavior.

[0043] These steps not only need to be completed in one go, but their processes and logic also need to be solidified to achieve automated processing, ensuring that future data can automatically complete the corresponding cleaning, transformation and processing procedures.

[0044] To ensure the accuracy of the calculated risk indicators, we need to remove transactions not initiated by the customer from the raw transaction data, such as reversal transactions and interest settlement transactions. This is achieved by analyzing three dimensions: transaction type, channel, and product type. For example, if a transaction is not initiated by the customer (such as an internal adjustment within the bank's system), it will not be included in the risk assessment model, thus avoiding interference with the final assessment results.

[0045] To accurately assess account risk and customer value, the model utilizes various types of data, primarily including the following five categories: Account and customer basic information data: including account status, account opening date, etc.

[0046] Customer value data: Reflects metrics such as customer contribution and loyalty.

[0047] Account fund transaction data: records of financial activities such as deposits and withdrawals.

[0048] Account behavior data: covering customer behavior such as login frequency and types of services used.

[0049] External data: Information from third parties, such as credit scores.

[0050] Each type of data contains several core fields that are crucial for building accurate assessment models. This approach enables a more comprehensive understanding and prediction of an account's risk profile and its holder's value potential.

[0051] Transactions unrelated to risk are excluded to avoid interfering with the calculation of risk indicators. These mainly include transactions not initiated by the customer. The exclusions are made based on three dimensions: transaction type, channel, and product type. Examples of the relevant logic are shown in Table 1 below.

[0052] Table 1. Relevant Logic type Filtering logic Transaction type Reversal trades (trade status: reversed, reversed) Bank fee transactions, interest settlement transactions Entrusted payment (collection) transactions Refund transactions channel Rights and interests platform Fund sales platform Fiscal payment system Wealth Management Platform product Personal time deposits Internal liquidation team and internal accounts Purchase products (financial products, insurance products, time deposits, loans, etc.) S120. Based on the preprocessing results, the account is classified as active or inactive according to its activity level to obtain the account type.

[0053] In this embodiment, account type refers to the process of classifying an account as an active or inactive account based on its activity level. Specifically, if an account's cumulative transaction amount reaches or exceeds a specific threshold on the statistical date and meets certain transaction count requirements, it is classified as an active account; otherwise, it is considered an inactive account.

[0054] Specifically, for example: an account is considered an active account if it has transactions on a given day, with a cumulative transaction amount of 1,000 yuan or more, at least 3 outgoing transactions, and at least 1 incoming transaction; conversely, an account that does not meet the criteria for an active account is considered an inactive account. The criteria for determining activity level may be adjusted based on actual business conditions.

[0055] To implement more precise and reasonable differentiated dynamic management of different bank accounts, it is necessary to first systematically review these accounts. Based on their attributes and usage characteristics, they can be categorized to facilitate targeted risk management and control measures.

[0056] For example, when involved in illegal activities, criminals often prepare accounts through so-called "credit card nurturing" before committing crimes. These accounts tend to have little or no activity, or only small, exploratory transactions, thus exhibiting low activity. Conversely, suspicious accounts used for criminal activities often involve large inflows and outflows of funds, showing high activity. Given the significantly different risk characteristics between credit card nurturing accounts and suspicious accounts used for criminal activities, they can be divided into two main categories based on their activity level: active accounts and inactive accounts.

[0057] An active account is defined as an account that has had transactions on a specific date (such as the statistical date) and meets the following conditions: The total transaction amount for the day reached or exceeded 1,000 yuan; The cumulative number of transactions processed must be no less than 3. There must be at least one transaction credited to your account.

[0058] Please note that specific numerical standards should be determined based on detailed data analysis; the examples above are for illustrative purposes only.

[0059] In contrast, inactive accounts refer to those that, although they have active transactions or non-funding transactions on the same day, fail to meet the above criteria for active accounts. Furthermore, if an account triggers an anti-fraud alert mechanism, even if its transaction volume or amount is small, it should be classified as an inactive account, as this may indicate potential risk factors.

[0060] By using this activity-based classification method, financial institutions can more effectively identify and monitor accounts with different risk types, thereby developing more scientific and reasonable control strategies. This approach not only helps improve the efficiency of risk management but also better protects customer assets.

[0061] S130. Based on the account type, the corresponding risk rating model is used to assess the risk level and identify potential risks to obtain the account risk.

[0062] In this embodiment, account risk refers to a quantitative value or level calculated using a specific algorithm or model based on a series of risk indicators, transaction behaviors, customer credit information, and other multi-dimensional data. This quantified risk value is used to measure the potential financial risks an account may be involved in within a certain period, such as the likelihood of money laundering, fraud, or other illegal activities. The level of account risk determines the level of monitoring and management measures that banks or other financial institutions need to take for that account.

[0063] In one embodiment, please refer to Figure 3 The above step S130 may include steps S131 to S134.

[0064] S131. The preprocessing results are processed using risk indicators to obtain the processed results.

[0065] In this embodiment, the processing result refers to the result obtained after in-depth analysis and processing of preprocessed data (such as basic account information, transaction records, behavioral patterns, etc.). This step typically involves using data mining techniques, machine learning models, and other methods to extract key indicators or features that can reflect the potential risks of the account. For example, abnormal behavior can be identified by analyzing factors such as the account's fund flow, transaction frequency, and transaction amount trends, and then converted into specific scores or values. These scores or values ​​will serve as inputs for subsequent risk assessment models.

[0066] At this stage, the system further processes and analyzes the collected data to extract key indicators that can be used for risk assessment. These indicators may include, but are not limited to, account activity frequency, transaction amount, transaction type, geographical location, and device usage. Through data mining and machine learning techniques, the system calculates a numerical value or score reflecting the potential risk of an account or transaction—the "processed result." This process provides input for subsequent risk assessment models.

[0067] S132. Determine the corresponding risk assessment model based on the account type.

[0068] Different account types (such as active and inactive accounts) may face different types of risks, thus requiring different risk assessment models. For example, active accounts may focus more on recent trading behavior and its trends, while inactive accounts may consider factors such as basic information and customer credit. This step involves selecting the most suitable risk assessment model based on the specific account type.

[0069] S133. Input the processing results (account classification) into the risk assessment model to identify potential risks and obtain the risk assessment results for this period.

[0070] In this embodiment, the current risk assessment result is the final conclusion obtained after further analysis of the processed results using a selected risk assessment model. This conclusion includes, but is not limited to, the account's risk level (e.g., risky, suspicious, watchful, normal, trustworthy), and specific risk management recommendations for the account. The assessment result not only reflects the overall risk status of the account in the current period but also considers historical risk rating results to more accurately update the account's risk classification. The risk assessment result provides decision support for financial institutions, helping them determine whether additional review, transaction restrictions, or fund freezing measures are necessary for the account.

[0071] In this step, the previously obtained processing results are input into the selected risk assessment model. The model will employ various strategies and techniques, such as list-based strategies, gambling / fraud detection strategies, and money laundering detection strategies, to conduct a comprehensive risk assessment of the account. During the assessment, not only will potential risk points be identified, but the extent to which these risks affect account security will also be evaluated. The final output is a specific assessment result, representing the account's risk level for the current period.

[0072] S134. Compare the current risk assessment results with the historical risk rating results, and update the account risk classification and grading according to the set logic to obtain the account risk.

[0073] Finally, the system will compare the newly derived risk assessment results for the current period with the account's historical risk rating results. Based on pre-defined logical rules, it will determine whether the account's risk classification needs adjustment. If the new assessment results indicate a higher risk, the account's risk level may need to be updated to allow for stricter monitoring measures or other risk management actions. Conversely, if the assessment results indicate a reduced risk, control measures for the account may be appropriately relaxed.

[0074] The entire process is a dynamic cycle designed to ensure the accuracy and timeliness of account risk assessment, thereby effectively preventing various financial risks.

[0075] Account risk rating first screens the accounts requiring rating for the current period, then processes the risk indicators needed for the corresponding model. This process includes classifying active and inactive accounts using specific risk assessment models, executing appropriate risk rating models based on their characteristics, and finally outputting the account's risk rating for the current period. Finally, a decision is made based on the account classification and grading strategy to update the current period's rating in the account risk classification and grading system.

[0076] For active account risk rating, this refers to a comprehensive risk assessment of accounts with recent transaction activity, considering their recent activity, transaction behavior, and risk characteristics to determine their risk level. Typically, account risk ratings are divided into five levels: Risky, Suspicious, Watchlist, Normal, and Trustworthy. The risk strategy process begins with an initial screening using list-based strategies. If an account or transaction matches a list-based strategy, it is directly classified as having a specific risk level (e.g., High Risk), and the assessment process ends. Accounts that do not match a list-based strategy proceed to the risk-based strategy assessment stage, employing various strategies such as gambling / fraud detection and money laundering detection for in-depth analysis. After the comprehensive assessment of the above steps, the risk level of the account or transaction is determined. The entire process aims to comprehensively and accurately identify and assess various risks, thereby taking corresponding control measures.

[0077] On the other hand, inactive account risk rating refers to a comprehensive risk assessment of accounts that have not had any recent transactions or have had transactions but do not meet the criteria for active accounts. This assessment considers factors such as basic information, customer credit, and customer behavior to determine their risk level. Similar to active accounts, inactive accounts also follow a list-based and risk-based assessment process. In the risk-based strategy phase, special attention is paid to analyzing unusual probing behavior, abnormal operational behavior, and behavioral correlation analysis to identify potential risk patterns. After a comprehensive assessment using these strategies, if suspicious behavior is found, the account is classified as suspicious; if no abnormalities are found, it is classified as normal. This multi-layered assessment mechanism helps improve the accuracy of risk identification and effectively prevents potential risk events from occurring.

[0078] Account classification and grading involves comparing the current risk rating status of an account with existing rating statuses within the system, and generating a latest risk rating and grading based on pre-defined scenario-based judgment logic. This process ensures that the account risk management system can be dynamically adjusted to reflect changes in account risk in a timely manner, enabling financial institutions to take appropriate management measures. Overall, this multi-layered assessment system provides a solid foundation for financial risk management and control.

[0079] The above-mentioned setting strategies are shown in Table 2.

[0080] Table 2. Setting Strategies Strategy Category Strategy subcategories How the strategy works List Strategy List Strategy By collecting and organizing existing whitelists or risk lists, classifying them according to their source and risk level, and performing matching, filtering, dynamic updates, and risk grading on new data and related party information, the system can quickly identify and manage risks while ensuring compliance requirements, thus improving the efficiency and accuracy of risk assessment. Furthermore, the list strategy is flexibly configurable to adapt to the risk management needs of different business scenarios. For example, inclusion and exclusion criteria for the list, as well as the basis for risk level classification, can be customized according to business rules, ensuring the targeted and effective nature of risk management. Simultaneously, the strategy supports retrospective analysis of historical data, helping managers gain a deeper understanding of risk trends and providing strong support for future risk decisions. Risk-based strategies Strategies for Unusual Probing Behaviors This strategy relies on the collection and analysis of diverse data, encompassing customer behavior data, transaction records, and other information. By employing rule-based detection, statistical methods, and machine learning algorithms (for example, rule-based detection can set specific thresholds for abnormal behavior, such as frequent login failures or abnormal transaction amounts within a short period), it conducts in-depth analysis and identification to uncover potential probing risk behaviors, such as small-scale, low-risk attempts. Such behaviors help identify abnormal patterns and allow for risk assessment of identified anomalies to determine their potential impact and severity. By analyzing the results of these attempts, the strategy can be adjusted to improve the accuracy and efficiency of detection. Anomaly probing behavior strategies are widely used in various fields, such as cybersecurity, where probing detection identifies potential attacks, and finance, where analyzing transaction behavior identifies fraudulent activities. Abnormal operation behavior strategy The principle behind abnormal operational behavior strategies is primarily based on the monitoring, analysis, and evaluation of customer or system operational behavior. Potential risks are identified by recognizing behaviors that deviate from normal patterns. For example, under regulatory compliance, detailed logs of customer or system operations are processed. These logs include raw data such as operation time, operation type, operation object, operation result, and IP address. After cleaning, filtering, and normalization, historical data is analyzed to establish normal behavioral patterns for customers or systems. Machine learning algorithms (such as clustering and classification) or deep learning models (such as CNN and RNN) are used to model and analyze operational behavior, identify abnormal patterns, and assess the risk based on the severity and frequency of detected abnormal behaviors. Due diligence risk identification strategy Due diligence risk identification strategies primarily rely on expert knowledge and data analysis to conduct comprehensive and multi-faceted risk identification. During communication and exchanges with target individuals, firsthand information is obtained through interviews, questionnaires, and other methods to ensure the authenticity and accuracy of the information. Particular attention is paid to identifying and assessing potential risks, and a feedback mechanism is established to ensure timely feedback on issues discovered during the investigation. Based on the feedback, advanced data analysis tools and models are used to conduct in-depth mining and detailed analysis of the collected data to further reveal potential risk points and their possible impact. The investigation strategy is dynamically adjusted to improve the accuracy and credibility of potential risk identification. Abnormal Change Behavior Strategy The operational mechanism of anomaly change behavior strategies is primarily based on the monitoring, analysis, and evaluation of behavioral data. Its purpose is to reveal potential risks by identifying behavioral changes that deviate from normal patterns. Implementing anomaly change behavior strategies first requires identifying patterns of normal behavior, a process typically accomplished by analyzing historical data. A baseline of normal behavior forms the basis for identifying anomalous behavior, enabling the system to distinguish between normal operations and potentially risky behaviors. For example, through machine learning and deep learning algorithms, the system can learn from large amounts of historical data to identify normal and anomalous behaviors. Furthermore, anomaly change behavior strategies need to be dynamically adjusted based on actual circumstances. For instance, when new anomalous patterns are discovered, the model can be updated or thresholds adjusted. In addition, a feedback mechanism feeds the detected anomalous behavior results back to the system to optimize the detection model. Therefore, anomaly change behavior strategies can effectively identify and manage potential risks. Abnormal trading behavior strategy The abnormal transaction behavior strategy is based on modeling normal transaction behavior and uses machine learning, statistical analysis, and other methods to identify transaction behaviors that deviate from normal patterns, monitoring high-frequency trading, fraudulent transactions, and other abnormal transaction behaviors in real time. The model and rules can be flexibly adjusted according to different business scenarios and risk preferences. This model is widely used in commercial banks and financial institutions, e-commerce platforms, and financial markets. For example, e-commerce platforms monitor abnormal transaction behaviors such as account activity, payment anomalies, and transaction similarity. Behavioral association mining strategy Behavioral association mining is a data mining technique that aims to reveal interesting connections or correlations between items in a dataset. In the field of risk behavior association mining, potential risk patterns can be identified by analyzing the association rules between different behaviors. The core principle of behavioral association mining strategies is to collect data related to customer behavior, transaction records, system operations, network traffic, etc., while ensuring data privacy and compliance. This data may come from various sources, including log files, databases, and sensors. After preprocessing steps such as cleaning, noise reduction, and normalization, the collected data is further used to extract features related to risk behaviors. For example, statistical analysis or machine learning techniques can be used to select the features most valuable for risk identification. Using methods such as information gain and chi-square tests to filter features can reveal risk patterns hidden in large amounts of data. The identified risk patterns and their assessment results are fed back to the system, helping to optimize models and strategies. By continuously updating data and models, the system can adapt to new risk patterns and behavioral changes. Focus Strategy Focus Strategy The core of a risk focus strategy lies in the continuous and meticulous monitoring of risks that cannot be completely avoided or effectively controlled, ensuring that risks can be quickly identified and addressed promptly and effectively when they first emerge. This strategy is not only an organic combination of risk appetite and risk control measures, but also a concrete manifestation of an enterprise's attitude towards specific risks and its choice of risk response strategies. Its main purpose is to appropriately manage and control risks that are difficult to completely eradicate or for which aggressive measures cannot be taken due to various factors, while maintaining a high level of attention and vigilance so that when risks truly manifest, a series of robust and proactive response measures can be quickly initiated and implemented, thereby minimizing the negative impact of risks on the enterprise's operations and development. A risk focus strategy is a dynamic and continuous process. Enterprises need to constantly adjust their risk focus and response measures according to changes in the internal and external environment to ensure the effectiveness and adaptability of risk management. Trusted Strategy Trusted Strategy The Trusted Strategy focuses on prioritizing the identification of high-quality accounts based on trustworthy characteristics to optimize strategy execution and decision-making processes. Generally, accounts with higher trustworthiness are considered more reliable based on trustworthy characteristics and are therefore given priority in the decision-making process. The Trusted Strategy needs to dynamically adjust trustworthy characteristics according to actual circumstances. This strategy can effectively prioritize the identification of high-quality accounts, improve the accuracy and efficiency of decision-making, and prevent high-quality accounts from being subject to strict spending limits. Therefore, the set strategies include list strategy, abnormal probing behavior strategy, abnormal operation behavior strategy, due diligence risk identification strategy, abnormal change behavior strategy, abnormal transaction behavior strategy, behavior correlation mining strategy, attention strategy, and trust strategy. The list strategy is to quickly identify and manage risks by matching and updating whitelists or risk lists, ensuring compliance and supporting flexible configuration to adapt to different business needs; The abnormal probing behavior strategy uses rules, statistical methods, and machine learning techniques to analyze customer behavior and transaction records in order to discover potential probing risk behaviors and assess the risk situation. Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal trading behavior strategies are based on modeling normal trading behavior and use methods such as machine learning and statistical analysis to identify trading behaviors that deviate from normal patterns and assess the risk situation. The due diligence risk identification strategy combines expert knowledge and data analysis tools, and collects information through interviews, questionnaires and other methods to identify potential risks and assess the risk situation. The abnormal behavior change strategy involves monitoring behavioral data and analyzing historical data, and establishing a baseline of normal behavior to identify abnormal behaviors that deviate from the normal pattern and assess the risk situation. Behavioral association mining strategy uses data mining techniques to analyze data from multiple sources, reveal the association rules between different behaviors to identify potential risk patterns and assess risk situations; The focus strategy involves continuous monitoring of risks that cannot be completely avoided or effectively controlled, so that swift action can be taken when the risks first emerge to reduce their impact on the business. Trustworthy strategies prioritize identifying high-quality accounts based on trustworthy characteristics, improving the accuracy and efficiency of decision-making, avoiding strict spending limits on these high-quality accounts, and can be dynamically adjusted according to actual circumstances.

[0081] S140. Analyze the customer's assets and transaction frequency indicators based on the preprocessing results, calculate the customer value score, and classify the value levels to obtain the customer value level.

[0082] In this embodiment, customer value level refers to the process of dividing customers into three or more value levels: high, medium, and low, based on a comprehensive value score calculated by analyzing multiple dimensions such as customer assets and transaction frequency.

[0083] In one embodiment, please refer to Figure 4 The above-mentioned step S140 may include steps S141 to S145.

[0084] S141. Collect relevant information from multiple dimensions such as customer assets, reputation, and account activity on the preprocessed results.

[0085] First, relevant information is collected from multiple dimensions, including the customer's assets, credit history, and account activity. This information includes, but is not limited to, the customer's asset status (such as monthly average daily AUM or current demand deposit balance), credit rating, and account activity level (such as the number of days in history where non-counter withdrawals were close to the limit). The main purpose of this stage is to comprehensively understand and quantify the customer's behavioral characteristics and their potential value to the bank.

[0086] S142. Process the relevant information, quantify and stratify it according to the characteristics that measure customer value, so as to obtain indicators.

[0087] Next, the collected information is processed, and indicators are quantified and stratified based on various characteristics that measure customer value. For example, customers are divided into different tiers based on their average daily AUM value over the past month; or the recent activity level of an account is assessed based on the average monthly number of historical withdrawal transactions, and stratified accordingly. In this way, the impact of each indicator on customer value can be more accurately reflected.

[0088] S143. Assign weights to each indicator based on its impact on customer value.

[0089] Then, based on the degree of impact of each indicator on customer value, an appropriate weight is assigned to each indicator. For example, indicators that better reflect the long-term stability and profitability of customers (such as monthly average daily AUM) may be given higher weights. This process needs to be carried out in conjunction with specific business needs and model design principles.

[0090] S144. Calculate the customer's overall value score using a weighted summation method.

[0091] Using the weights determined in the previous step, a weighted summation method is employed to calculate the customer's overall value score. This means multiplying the scores of all relevant indicators by their corresponding weights and then summing them to obtain a total score. This score reflects the customer's overall value level relative to other customers.

[0092] S145. Based on the preset scoring range of the comprehensive value scoring criteria, customers are divided into three or more value levels: high, medium, and low, to obtain customer value levels.

[0093] Finally, based on pre-defined scoring ranges (e.g., [100+) represents high value, [50,100) represents medium value, and [0,50) represents low value), customers are categorized into high, medium, and low value levels. This step helps financial institutions develop differentiated service strategies and risk management measures based on customers' actual value contributions, thereby improving customer satisfaction and service efficiency.

[0094] In summary, this process, through a systematic approach to comprehensively assess customers, not only accurately identifies high-risk customers but also provides valuable customers with opportunities to enhance their customer experience, thereby promoting the overall development of financial institutions.

[0095] Customer value rating is primarily defined through characteristics such as customer asset information, credit information, customer contribution, account activity, transaction frequency, and credit limit usage. This comprehensive assessment determines a customer's value contribution to the bank and categorizes them into high-value, medium-value, low-value, or multiple value levels. The overall customer value assessment can be adjusted and updated promptly based on data.

[0096] The model design can be based on the specific business indicators input into the model, and weight calculation can be performed using a scorecard-like approach.

[0097] Some example features are shown in Table 3 below.

[0098] Table 3. Characteristics Evaluation Dimensions Measurement indicators Measurement Description Customer contribution Customer level or card type Measuring customer segmentation Average AUM (Assets) per Month Measure the customer's average daily AUM over the past month and segment them. Current AUM or Current Demand Deposit Balance Measure and stratify customers' recent demand deposit balances. Customer activity The number of days in history where non-counter withdrawals were close to the limit Assess and stratify customers' recent demand for non-counter credit limits. Historical monthly or daily average number of withdrawal transactions (average daily number of transactions from the previous month) Measure the average size of recent withdrawal transactions in the account and stratify them. Each feature in the customer value rating model has a different weight. A corresponding weight score is assigned to each feature input into the model, and the total customer value score is calculated comprehensively. Different scores correspond to different customer value levels. An example is shown in Table 4 below.

[0099] Table 4. Customer Value Levels Customer Value Rating Customer Value Level [100+） High value [50,100） Medium Value [0,50） low value S150. Automatically adjust the account transaction limit based on the account risk and the customer value level to obtain the adjustment result.

[0100] In this embodiment, the adjustment result refers to the precise and flexible adjustment of account transaction limits based on the account's risk level and the customer's value level through an automated dynamic management scheme. Specifically, this process first determines the most suitable fixed or dynamic limit reduction or increase scheme based on different combinations of account risk and customer value level, so as to achieve effective limit control for accounts with different risks and values. While effectively preventing risks, it provides matching limits for customers of different levels, optimizing customer experience. Furthermore, when a real-time change in account risk level or customer value level is detected, the system can automatically trigger the corresponding limit adjustment process. Combining the pre-set matrix limit adjustment strategy and the current account limit status, it automatically generates and executes specific operations to reduce or increase the limit, ultimately obtaining an adjustment result that meets actual business needs.

[0101] In one embodiment, please refer to Figure 5 The above-mentioned step S150 may include steps S151 to S153.

[0102] S151. Based on the account risk and the customer value level, determine different fixed or dynamic credit limit reduction or increase plans to effectively prevent risks while providing matching credit limits for customers of different levels and optimizing customer experience.

[0103] Develop specific fixed or dynamic limit reduction plans based on account risk and customer value levels. This step focuses on how to control potential risks by limiting transaction limits for high-risk, low-value accounts.

[0104] S152. Design tiered credit limit increase rules based on the account risk and the customer value level to provide corresponding credit limit increases for accounts of different levels.

[0105] Based on account risk and customer value levels, a detailed tiered credit limit increase rule is designed. This step aims to ensure that there are appropriate and reasonable credit limit increase mechanisms for accounts of different levels, thereby incentivizing customer behavior and promoting positive business interactions.

[0106] S153. When account risk or customer value level changes in real time, the credit limit adjustment process is automatically triggered. Based on the defined matrix credit limit adjustment scheme and strategy, combined with the current account credit limit, the operation of reducing or increasing the credit limit is automatically generated and executed to obtain the adjustment result.

[0107] Once account risk or customer value level changes, the system will automatically initiate a credit limit adjustment process. Utilizing pre-defined matrix adjustment schemes and strategies, combined with the specific circumstances of the current account, the system can instantly generate and execute corresponding credit limit adjustments. Whether lowering or raising the account limit, the aim is to maximize adaptability to the current business scenario and risk management requirements. This automated process ensures the timeliness and accuracy of limit adjustments, helping to maintain good customer relationships while minimizing risk exposure.

[0108] In this embodiment, a dynamic credit limit adjustment strategy is formulated based on a two-dimensional matrix of account risk classification and customer value levels. The adjustment approach aims to limit or increase the credit limits of all accounts under a more reasonable strategy, taking into account actual business scenarios. For example, high-risk, low-value accounts can be strictly limited to control risk, while low-risk, high-value accounts can have their limits increased to improve customer experience. Account limit strategies can also adopt methods such as fixed low limits, fixed specific limits, and dynamic reduction based on the current credit limit.

[0109] The trigger strategy for account limit adjustments can be triggered in real time by changes in account risk classification or customer value level. It can adapt to different customer groups, applicable scenarios, and differentiated risk control strategies, adapting the corresponding account control limit strategy in real time. After obtaining the account's risk level and customer value rating, the most crucial step in this invention is implemented: a limit strategy matrix scheme based on the account risk level and customer value level, along with the current limit adjustment strategy. This matrix scheme and strategy automatically outputs the result of reducing or increasing the account limit; finally, the updated limit is synchronized to the business system.

[0110] By combining the risk level and value level of the account, a matrix-based dynamic limit adjustment strategy is formed, as shown in Table 5 below. An example of the dynamic account limit management scheme is shown in Table 6.

[0111] Table 5. Matrix Quota Dynamic Adjustment Strategy

[0112] Table 6. Dynamic Management Scheme for Account Limits

[0113] For example, in the batch data from the previous day's model, there were two accounts, A and B, for which risk ratings were conducted. Both accounts changed from "normal" to "suspicious," indicating an increase in account risk level. Account A had a high value, while account B had a low value. Based on the pre-set dynamic limit adjustment strategy, the limit control limit output from the previous day's strategy was 50,000 yuan for account A and 20,000 yuan for account B. In yesterday's batch data from the model, the risk level of account A was raised to "risk" and the risk level of account B was lowered to "normal". The value corresponding to account A remains high value and the value corresponding to account B remains low value. According to the set dynamic adjustment strategy, the control limit output by yesterday's limit strategy is 10,000 yuan for account A and 40,000 yuan for account B.

[0114] S160. Synchronize the adjustment results to the relevant business systems.

[0115] The latest credit limit adjustment results will be automatically pushed to various business systems and channels.

[0116] Based on business needs and target strategies, relevant field data is collected from different data platforms. The raw data is then filtered and processed, taking into account actual business data quality and metrics. Using accounts and customers as core dimensions, the collected data is cleaned, transformed, and aggregated to generate basic data and metrics suitable for model input.

[0117] A comprehensive review of bank accounts was conducted, and they were classified according to established plans and statistical standards.

[0118] The system applies a corresponding risk rating model to each account category to determine the risk level of each account. Then, it automatically updates the risk classification results of the accounts according to preset logic.

[0119] Run the customer value assessment model and output the final customer value rating.

[0120] By combining the risk classification and grading of accounts with the value assessment results of customers, a pre-defined dynamic credit limit adjustment strategy matrix is ​​applied to automatically generate the latest credit limit adjustment plan for all accounts.

[0121] The latest credit limit adjustment results are automatically synchronized to various business systems and channels to ensure the timeliness and accuracy of business operations.

[0122] This embodiment integrates the results of account risk rating and customer value rating models to achieve a comprehensive assessment of account risk and value, providing a solid foundation for differentiated credit limit adaptation. Employing a real-time dynamic credit limit adjustment matrix strategy, it intelligently adjusts credit limits based on real-time changes in account risk and value, improving the accuracy and efficiency of credit limit management. While ensuring normal business operations, it fully considers the inherent risk characteristics of accounts, achieving intelligent differentiated credit limit adaptation and effectively balancing risk management and customer service experience. The entire process is highly automated, from data collection to the delivery of credit limit adjustment results, reducing manual intervention, improving processing speed and accuracy, and ensuring immediate responsiveness in business operations.

[0123] The method in this embodiment aims to monitor and analyze customer account transaction behavior in real time, and to conduct a comprehensive evaluation by combining multi-dimensional information such as cardholder risk characteristics, transaction patterns, and customer value. Based on this comprehensive evaluation system and dynamic credit limit adjustment matrix strategy, the system can automatically and intelligently adjust the account's transaction limits in response to real-time updates on changes in account risk status and value.

[0124] Through continuous monitoring and in-depth analysis of customer account transactions, the system can promptly detect any abnormal transaction behavior or potential risk signals. It integrates various factors such as cardholder risk characteristics, transaction behavior, and customer value to form a comprehensive customer profile, providing a basis for personalized services. Based on changes in account risk and value information, a pre-set dynamic credit limit adjustment matrix strategy is employed to achieve intelligent and automated adjustments to account transaction limits. While effectively preventing financial risks such as fraud and money laundering, it ensures that customers' normal card usage needs are met, thereby achieving the optimal balance between fund security and transaction convenience.

[0125] The implementation of this method not only enhances banks' ability to manage account risks but also improves customer service experience. Through precise risk control and flexible credit limit adjustment mechanisms, it safeguards customer funds while promoting a convenient transaction environment. This represents a significant advancement in banking business processes, contributing to improved overall operational efficiency and service quality. It not only enhances the scientific rigor and accuracy of account management but also optimizes customer service experience while effectively preventing potential risks.

[0126] The aforementioned dynamic account limit adjustment method acquires and preprocesses account-related information, classifies accounts into active or inactive types based on their activity, applies corresponding risk rating models to different types to identify potential risks, analyzes customer assets and transaction frequency indicators to calculate customer value scores, and finally automatically adjusts account transaction limits based on account risk and customer value level, synchronizing the adjustment results to relevant business systems. This achieves a more intelligent, dynamic, and highly adaptable account limit adjustment mechanism. This method solves the problems of existing technologies, such as lack of real-time performance, insufficient flexibility, difficulty in effectively distinguishing between normal and potentially risky customers, and misjudgment of high-value customers. It achieves the goal of meeting the needs of different customers while ensuring security, improving the accuracy of risk management and service quality. Through this comprehensive assessment and automated adjustment approach, it not only enhances the ability to prevent fraudulent and other risky behaviors but also ensures a better service experience for high-value customers.

[0127] Figure 6 This is a schematic block diagram of an account limit dynamic adjustment device 300 provided in an embodiment of the present invention. Figure 6 As shown, corresponding to the above-described method for dynamically adjusting account limits, the present invention also provides an account limit dynamic adjustment device 300. This device 300 includes a unit for executing the above-described method for dynamically adjusting account limits, and can be configured in a server. Specifically, please refer to... Figure 6 The account limit dynamic adjustment device 300 includes an acquisition unit 301, a classification unit 302, a risk assessment unit 303, a level determination unit 304, an adjustment unit 305, and a synchronization unit 306.

[0128] The system comprises the following components: an acquisition unit 301, which acquires and cleans / filters account-related information to obtain preprocessing results; a classification unit 302, which classifies accounts as active or inactive based on their activity according to the preprocessing results to obtain account type; a risk assessment unit 303, which assesses account risk level using a corresponding risk rating model based on account type and identifies potential risks to obtain account risk; a level determination unit 304, which analyzes customer assets and transaction frequency indicators based on the preprocessing results, calculates customer value score, and classifies value levels to obtain customer value level; an adjustment unit 305, which automatically adjusts account transaction limits based on account risk and customer value level to obtain adjustment results; and a synchronization unit 306, which synchronizes the adjustment results to relevant business systems.

[0129] In one embodiment, the acquisition unit 301 is used to acquire account-related information required for account risk and customer value assessment, and to obtain preprocessed results by cleaning, transforming and processing the account-related information and filtering out transactions not initiated by the customer.

[0130] In one embodiment, the risk assessment unit 303 includes: The processing subunit is used to process the preprocessing results with risk indicators to obtain the processing result; the determination subunit is used to determine the corresponding risk assessment model based on the account type; the assessment subunit is used to input the processing result into the risk assessment model and identify potential risks to obtain the assessment result; the comparison subunit is used to compare the current period risk assessment result with the historical risk rating result, and update the account risk classification level according to the set logic to obtain the account risk.

[0131] In one embodiment, the level determination unit 304 includes: The system comprises the following sub-units: a collection sub-unit, which collects relevant information from multiple dimensions, including customer assets, reputation, and account activity, based on the preprocessing results; a processing sub-unit, which processes the relevant information, quantifies and stratifies each measurement indicator to obtain the indicators; a weighting sub-unit, which assigns weights to each indicator based on its impact on customer value; a summation sub-unit, which calculates the customer's comprehensive value score using a weighted summation method; and a segmentation sub-unit, which divides the customer into three or more value levels (high, medium, and low) based on the comprehensive value score and a preset scoring range to obtain the customer value level.

[0132] In one embodiment, the adjustment unit 305 includes: The scheme determination subunit is used to determine different fixed or dynamic credit limit reduction schemes based on the account risk and the customer value level, and to implement credit limit control for high-risk, low-value accounts. The credit limit increase determination subunit is used to design tiered credit limit increase rules based on the account risk and the customer value level, and to provide corresponding credit limit increases for accounts of different levels. The triggering subunit is used to automatically trigger the credit limit adjustment process when the account risk or customer value level changes in real time. Based on the defined matrix credit limit adjustment scheme and strategy, and combined with the current account credit limit situation, it automatically generates and executes the operation of reducing or increasing the credit limit to obtain the adjustment result.

[0133] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned account limit dynamic adjustment device 300 and each unit can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0134] The aforementioned account limit dynamic adjustment device 300 can be implemented as a computer program, which can, for example... Figure 7 It runs on the computer device shown.

[0135] Please see Figure 7 , Figure 7 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0136] See Figure 7 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0137] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a method for dynamically adjusting account limits.

[0138] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0139] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a method for dynamically adjusting the account limit.

[0140] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0141] The processor 502 is used to run a computer program 5032 stored in the memory to perform the following steps: The process involves: acquiring and cleaning / filtering account-related information to obtain preprocessing results; classifying accounts as active or inactive based on their activity levels to determine account types; assessing account risk using a corresponding risk rating model to identify potential risks; analyzing customer assets and transaction frequency indicators to calculate customer value scores and assigning value levels to determine customer value levels; automatically adjusting account transaction limits based on account risk and customer value levels to obtain adjustment results; and synchronizing these adjustment results to relevant business systems.

[0142] The term "active account" refers to an account that has active transactions on the same day, with a cumulative transaction amount of 1,000 yuan or more, at least 3 outgoing transactions, and at least 1 incoming transaction. "Inactive account" refers to an account that has no active transactions on the same day or has active transactions but does not meet the criteria for an active account. The criteria for determining activity level may be adjusted based on actual business conditions.

[0143] In one embodiment, when the processor 502 implements the step of obtaining account-related information and performing cleaning and filtering to obtain preprocessing results, it specifically implements the following steps: Obtain account-related information required for account risk and customer value assessment, cleanse, transform and process the account-related information, and filter out transactions not initiated by the customer to obtain preprocessed results.

[0144] In one embodiment, when the processor 502 implements the step of assessing the risk level based on the account type using the corresponding risk rating model and identifying potential risks to obtain the account risk, the processor 502 specifically implements the following steps: The preprocessing results are processed using risk indicators to obtain the processed results; a corresponding risk assessment model is determined based on the account type; the processed results are input into the risk assessment model to identify potential risks and assess their impact to obtain the assessment results; the current period risk assessment results are compared with historical risk rating results, and the account risk classification and grading are updated according to the set logic to obtain the account risk.

[0145] The set strategies include list strategy, abnormal probing behavior strategy, abnormal operation behavior strategy, due diligence risk identification strategy, abnormal change behavior strategy, abnormal transaction behavior strategy, behavior correlation mining strategy, attention strategy, and trust strategy.

[0146] The list strategy is to quickly identify and manage risks by matching and updating whitelists or risk lists, ensuring compliance and supporting flexible configuration to adapt to different business needs; Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal trading behavior strategies are based on modeling normal trading behavior and use methods such as machine learning and statistical analysis to identify trading behaviors that deviate from normal patterns and assess the risk situation. The due diligence risk identification strategy combines expert knowledge and data analysis tools, and collects information through interviews, questionnaires and other methods to identify potential risks and assess the risk situation. The abnormal behavior change strategy involves monitoring behavioral data and analyzing historical data, and establishing a baseline of normal behavior to identify abnormal behaviors that deviate from the normal pattern and assess the risk situation. Behavioral association mining strategy uses data mining techniques to analyze data from multiple sources, reveal the association rules between different behaviors to identify potential risk patterns and assess risk situations; The focus strategy involves continuous monitoring of risks that cannot be completely avoided or effectively controlled, so that swift action can be taken when the risks first emerge to reduce their impact on the business. Trustworthy strategies prioritize identifying high-quality accounts based on trustworthy characteristics, improving the accuracy and efficiency of decision-making, preventing these high-quality accounts from being subject to strict spending limits, and trustworthy strategies can be dynamically adjusted according to actual circumstances.

[0147] In one embodiment, when the processor 502 performs the step of analyzing the customer's assets and transaction frequency indicators based on the preprocessed results, calculating the customer value score, and classifying the value levels to obtain the customer value level, the processor 502 specifically implements the following steps: The preprocessing results collect relevant information from multiple dimensions, including customer assets, reputation, and account activity. This information is then processed, quantified, and stratified according to the characteristics used to measure customer value to obtain indicators. Weights are assigned to each indicator based on its impact on customer value. A weighted summation method is used to calculate the customer's overall value score. Based on the overall value score and a preset scoring range, customers are divided into three or more value levels: high, medium, and low, to obtain customer value levels.

[0148] In one embodiment, when implementing the step of automatically adjusting the account transaction limit based on the account risk and the customer value level to obtain the adjustment result, the processor 502 specifically implements the following steps: Based on the account risk and customer value level, different fixed or dynamic credit limit reduction schemes are determined to control the risk of high-risk, low-value accounts. Based on the account risk and customer value level, tiered credit limit increase rules are designed to provide corresponding credit limit increases for accounts of different levels. When the account risk or customer value level changes in real time, the credit limit adjustment process is automatically triggered. According to the defined matrix credit limit adjustment scheme and strategy, combined with the current account credit limit situation, the operation of reducing or increasing the credit limit is automatically generated and executed to obtain the adjustment result.

[0149] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0150] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0151] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform the following steps: The process involves: acquiring and cleaning / filtering account-related information to obtain preprocessing results; classifying accounts as active or inactive based on their activity levels to determine account types; assessing account risk using a corresponding risk rating model to identify potential risks; analyzing customer assets and transaction frequency indicators to calculate customer value scores and assigning value levels to determine customer value levels; automatically adjusting account transaction limits based on account risk and customer value levels to obtain adjustment results; and synchronizing these adjustment results to relevant business systems.

[0152] The term "active account" refers to an account that has active transactions on the same day, with a cumulative transaction amount of 1,000 yuan or more, at least 3 outgoing transactions, and at least 1 incoming transaction. "Inactive account" refers to an account that has no active transactions on the same day or has active transactions but does not meet the criteria for an active account. The criteria for determining activity level may be adjusted based on actual business conditions.

[0153] In one embodiment, when the processor executes the computer program to implement the step of obtaining account-related information and performing cleaning and filtering to obtain preprocessing results, it specifically implements the following steps: Obtain account-related information required for account risk and customer value assessment, cleanse, transform and process the account-related information, and filter out transactions not initiated by the customer to obtain preprocessed results.

[0154] In one embodiment, when the processor executes the computer program to perform the step of assessing the risk level based on the account type using the corresponding risk rating model and identifying potential risks, it specifically implements the following steps: The preprocessing results are processed using risk indicators to obtain a processed result; a corresponding risk assessment model is determined based on the account type; the processed result is input into the risk assessment model, and potential risks are identified and their impact is assessed to obtain an assessment result; the current period risk assessment result is compared with the historical risk rating result, and the account risk classification is updated according to the set logic to obtain the account risk.

[0155] The set strategies include list strategy, abnormal probing behavior strategy, abnormal operation behavior strategy, due diligence risk identification strategy, abnormal change behavior strategy, abnormal transaction behavior strategy, behavior correlation mining strategy, attention strategy, and trust strategy.

[0156] Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal trading behavior strategies are based on modeling normal trading behavior and use methods such as machine learning and statistical analysis to identify trading behaviors that deviate from normal patterns and assess the risk situation. The due diligence risk identification strategy combines expert knowledge and data analysis tools, and collects information through interviews, questionnaires and other methods to identify potential risks and assess the risk situation. The abnormal behavior change strategy involves monitoring behavioral data and analyzing historical data, and establishing a baseline of normal behavior to identify abnormal behaviors that deviate from the normal pattern and assess the risk situation. Behavioral association mining strategy uses data mining techniques to analyze data from multiple sources, reveal the association rules between different behaviors to identify potential risk patterns and assess risk situations; The focus strategy involves continuous monitoring of risks that cannot be completely avoided or effectively controlled, so that swift action can be taken when the risks first emerge to reduce their impact on the business. Trustworthy strategies prioritize identifying high-quality accounts based on trustworthy characteristics, improving the accuracy and efficiency of decision-making, avoiding strict spending limits on these high-quality accounts, and can be dynamically adjusted according to actual circumstances.

[0157] In one embodiment, when the processor executes the computer program to analyze indicators such as customer assets and transaction frequency, calculate customer value scores, and classify value levels to obtain customer value levels, the specific steps are as follows: The preprocessing results collect relevant information from multiple dimensions, including customer assets, reputation, and account activity. This information is then processed, quantified, and stratified according to the characteristics used to measure customer value to obtain indicators. Weights are assigned to each indicator based on its impact on customer value. A weighted summation method is used to calculate the customer's overall value score. Based on the overall value score and a preset scoring range, customers are divided into three or more value levels: high, medium, and low, to obtain customer value levels.

[0158] In one embodiment, when the processor executes the computer program to automatically adjust the account transaction limit based on the account risk and the customer value level to obtain the adjustment result, it specifically implements the following steps: Based on the account risk and customer value level, different fixed or dynamic credit limit reduction schemes are determined to control the risk of high-risk, low-value accounts. Based on the account risk and customer value level, tiered credit limit increase rules are designed to provide corresponding credit limit increases for accounts of different levels. When the account risk or customer value level changes in real time, the credit limit adjustment process is automatically triggered. According to the defined matrix credit limit adjustment scheme and strategy, combined with the current account credit limit situation, the operation of reducing or increasing the credit limit is automatically generated and executed to obtain the adjustment result.

[0159] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0160] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0161] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0162] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0163] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0164] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for dynamically adjusting account limits, characterized in that, include: Obtain account-related information and perform cleaning and filtering to obtain preprocessed results; Based on the preprocessing results, the accounts are classified as active or inactive according to their activity levels to obtain the account type; Based on the account type, the corresponding risk rating model is used to assess the risk level and identify potential risks in order to obtain the account risk. The preprocessing results are analyzed to determine the customer's assets and transaction frequency indicators, calculate the customer value score, and classify the customer into value levels to obtain the customer value level. The account transaction limit is automatically adjusted based on the account risk and the customer value level to obtain the adjustment result; The adjustment results will be synchronized to the relevant business systems.

2. The method for dynamically adjusting account limits according to claim 1, characterized in that, The process of obtaining account-related information and cleaning and filtering it to obtain preprocessed results includes: Obtain account-related information required for account risk and customer value assessment, cleanse, transform and process the account-related information, and filter out transactions not initiated by the customer to obtain preprocessed results.

3. The method for dynamically adjusting account limits according to claim 1, characterized in that, The activity level of an account is categorized based on factors such as the cumulative amount and frequency of transactions on a given day. For example, an account with transactions on a given day, where the cumulative transaction amount reaches or exceeds 1,000 yuan, with at least 3 outgoing transactions and at least 1 incoming transaction, is considered an active account; conversely, an account that does not meet the criteria for an active account is considered an inactive account. The criteria for determining activity level may be adjusted based on actual business needs.

4. The method for dynamically adjusting account limits according to claim 1, characterized in that, The risk level assessment based on the account type using the corresponding risk rating model, identifying potential risks, and obtaining account risk includes: The preprocessing results are then processed using risk indicators to obtain account classifications; Determine the appropriate risk assessment model based on the account type; Based on the account classification, the account is input into the corresponding risk assessment model to identify the potential risks of the account in order to obtain the risk assessment results for this period; By comparing the current assessment results with the historical risk rating results, the risk classification and grading of the account are updated according to the established logic to obtain the account risk.

5. The method for dynamically adjusting account limits according to claim 4, characterized in that, The set strategies include list strategy, abnormal probing behavior strategy, abnormal operation behavior strategy, abnormal transaction behavior strategy, due diligence risk identification strategy, abnormal change behavior strategy, behavior correlation mining strategy, attention strategy, and trust strategy. The list strategy is to quickly identify and manage risks by matching and updating whitelists or risk lists, ensuring compliance and supporting flexible configuration to adapt to different business needs; The abnormal probing behavior strategy uses rules, statistical methods, and machine learning techniques to analyze customer behavior and transaction records in order to discover potential probing risk behaviors and assess the risk situation. Abnormal operation behavior strategy is based on in-depth analysis of customer operation log data in the system, using machine learning or deep learning models to identify operational behaviors that deviate from the normal pattern and assess the risk situation; Abnormal trading behavior strategies are based on modeling normal trading behavior and use methods such as machine learning and statistical analysis to identify trading behaviors that deviate from normal patterns and assess the risk situation. The due diligence risk identification strategy combines expert knowledge and data analysis tools, and collects information through interviews, questionnaires and other methods to identify potential risks and assess the risk situation. The abnormal behavior change strategy involves monitoring behavioral data and analyzing historical data, and establishing a baseline of normal behavior to identify abnormal behaviors that deviate from the normal pattern and assess the risk situation. Behavioral association mining strategy uses data mining techniques to analyze data from multiple sources, uncover association rules between different behaviors to identify potential risk patterns, and assess risk situations; The focus strategy involves continuous monitoring of risks that cannot be completely avoided or effectively controlled, so that swift action can be taken when the risks first emerge to reduce their impact on the business. Trustworthy strategies prioritize identifying high-quality accounts based on trustworthy characteristics, improving the accuracy and efficiency of decision-making, avoiding strict spending limits on these high-quality accounts, and can be dynamically adjusted according to actual circumstances.

6. The method for dynamically adjusting account limits according to claim 1, characterized in that, Analyze customer assets and transaction frequency metrics to calculate customer value scores and classify them into value levels, including: Collect relevant information from multiple dimensions such as customer assets, credit, and account activity. The relevant information is processed, quantified and stratified according to the characteristics that measure customer value, to obtain indicators; Based on the degree of impact of the indicators on customer value, assign weights to each indicator; Calculate the customer's overall value score using a weighted summation method; Based on the comprehensive value score, customers are divided into three or more value levels: high, medium, and low, to obtain customer value levels.

7. The method for dynamically adjusting account limits according to claim 1, characterized in that, Based on the account risk and customer value level, a corresponding dynamic limit adjustment strategy is automatically matched to automatically adjust the account's transaction limits, including: Based on the account risk and the customer value level, different fixed or dynamic limit reduction schemes are determined, for example: implementing limit control risk for high-risk, low-value accounts; Based on the account risk and the customer value level, a tiered credit limit increase rule is designed to provide corresponding credit limit increases for accounts of different levels. For example, increasing the credit limit for low-risk, high-value accounts can improve the customer experience. When account risk or customer value level changes in real time, the credit limit adjustment process is automatically triggered. Based on the defined matrix credit limit adjustment plan and strategy, combined with the current account credit limit, the operation of reducing or increasing the credit limit is automatically generated and executed to obtain the adjustment result.

8. A device for dynamically adjusting account limits, characterized in that, include: The acquisition unit is used to acquire account-related information and perform cleaning and filtering to obtain preprocessed results; A classification unit is used to classify accounts as active or inactive based on the preprocessing results and their activity status to obtain the account type; The risk assessment unit is used to assess the risk level based on the account type using the corresponding risk rating model, identify potential risks, and obtain the account risk. The rating determination unit is used to analyze the customer's assets and transaction frequency indicators based on the preprocessing results, calculate the customer value score, and classify the value levels to obtain the customer value level. An adjustment unit is used to automatically adjust the account transaction limit based on the account risk and the customer value level to obtain the adjustment result; The synchronization unit is used to synchronize the adjustment results to the relevant business systems.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images