Modeling data service management method and system for digital intelligent financial ecology

Abstract

The invention discloses a modeling data service management method and system for digital intelligent financial ecology, and belongs to the technical field of data management. The method comprises the following steps: acquiring financial service events generated by a user on a digital intelligent financial ecological platform and timestamps when the financial service events are generated, constructing a financial service event chain and a financial service event chain matrix, and calculating event switching weights of the user under adjacent timestamps; calculating the weight fluctuation amplitude of the user between adjacent event switching weights based on the event switching weights; based on the weight fluctuation amplitude, calculating the accumulated weight fluctuation amplitude of the user under all timestamps; predicting the event switching trend of the user in the future based on the accumulated weight fluctuation amplitude; and presetting a threshold value, analyzing and performing unified management. According to the method, the time sequence event switching weight and fluctuation amplitude calculation are combined, the potential financial risk can be accurately warned, and the risk control and service strategy under the digital intelligent financial ecology is optimized, so that the intelligent level of financial management is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and specifically to a method and system for model-based data service management for the digital and intelligent financial ecosystem. Background Art

[0002] In recent years, with the rapid development of fintech, the digital and intelligent financial ecosystem (digital and intelligent finance) has gradually become an important mode of financial services; digital and intelligent finance has realized the intelligent upgrade of financial services through advanced technologies such as big data, artificial intelligence, blockchain, and cloud computing, and promoted personalized, real-time, and intelligent financial products and services. At present, the financial data management methods widely used in the market mainly rely on traditional database management systems, machine learning prediction models, and statistical analysis methods, mainly focusing on aspects such as user credit scoring, risk control, and personalized financial product recommendations. However, when dealing with high-frequency, heterogeneous, and complex user financial transaction data, these methods still have problems such as data isolation, difficulty in constructing event chains, and difficulty in predicting dynamic financial behaviors, and cannot comprehensively depict the financial behavior patterns of users.

[0003] Existing financial data management methods often have a static storage structure oriented to accounts and are difficult to adapt to the dynamic changes of user financial behaviors. Secondly, most of the current mainstream financial data modeling methods are based on static analysis of user historical transaction records, lacking modeling of the event switching characteristics in the time series dimension and being difficult to accurately predict the future financial behavior trends of users. In addition, when measuring the impact degree of user financial event switching, existing methods often rely on fixed weights or empirical weights based on statistics, without fully considering the comprehensive impact of event types, transaction amounts, and behavior switching patterns. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for model-based data service management for the digital and intelligent financial ecosystem to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the present invention provides the following technical solutions:

[0006] A model-based data service management method for the digital intelligence financial ecosystem. This method includes the following steps: Step S1: Obtain the financial service events generated by users on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the deposit events and loan events of users; Step S2: Construct the financial service event chain corresponding to a user at a single timestamp. The financial service event chain includes a single timestamp, the financial service event corresponding to the single timestamp, and the event amount of the financial service event; Construct a financial service event chain matrix; Step S3: Based on the financial service event chain matrix, calculate the event switching weight of the user at adjacent timestamps; Step S4: Based on the event switching weight, calculate the weight fluctuation amplitude between adjacent event switching weights; Based on the weight fluctuation amplitude, calculate the cumulative weight fluctuation amplitude of the user at all timestamps; Based on the cumulative weight fluctuation amplitude, predict the future event switching trend of the user; Preset a threshold, analyze and conduct unified management.

[0007] As a preferred solution of the model-based data service management method for the digital intelligence financial ecosystem described in the present invention, based on the digital intelligence financial ecosystem platform, obtain the behavior events of users. The behavior events include the financial service events generated by users on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the deposit events and loan events of users.

[0008] Denote the financial service event corresponding to the t-th timestamp as FSE t,i , where t ∈ [1, T], i ∈ [1, 2], T represents the total number of timestamps, i represents the index variable for distinguishing deposit events and loan events, and FSE t,1 indicates that the financial service event corresponding to the t-th timestamp is a deposit event, and FSE t,2 indicates that the financial service event corresponding to the t-th timestamp is a loan event.

[0009] As a preferred solution of the model-based data service management method for the digital intelligence financial ecosystem described in the present invention, based on the timestamp, construct the financial service event chain corresponding to the user at the t-th timestamp, denoted as {(TI t , FSE t,i , M(FSE t,i )) || i ∈ [1, 2]}; Obtain the financial service event chains corresponding to all T timestamps, and sort the timestamps according to the ascending time rule, construct a financial service event chain matrix, and use the financial service event chain as the row of the matrix, specifically as follows:

[0010]

[0011] where TI trepresents the t-th timestamp, and M(FSE t,i ) represents the event amount of the financial service event corresponding to the t-th timestamp.

[0012] As a preferred solution of the method for modeling data service management for the digital intelligence financial ecosystem described in the present invention, based on the financial service event chain matrix, calculate the event switching weight of the user under adjacent timestamps, and the calculation formula of the event switching weight is as follows:

[0013] If the financial service events under adjacent timestamps are the same, that is, FSE t,i = FSE t-1,i , then the event switching weight formula is as follows:

[0014]

[0015] Among them, W(t - 1→t) represents the event switching weight from the (t - 1)-th timestamp to the t-th timestamp, α represents the weight coefficient when the event type remains unchanged, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t - 1)-th timestamp.

[0016] If the financial service events under adjacent timestamps are different, that is, FSE t,i ≠FSE t-1,i , then the event switching weight formula is as follows:

[0017]

[0018] Among them, W(t - 1→t) represents the event switching weight from the (t - 1)-th timestamp to the t-th timestamp, β represents the weight coefficient when the event type changes, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t - 1)-th timestamp.

[0019] In the present invention, this formula means that when the event type switches: take the average value of the event amounts of two adjacent timestamps, and then multiply by the absolute value of the relative amplitude of the amount change. This not only considers the size of the amount, but also highlights the degree of amount change. When the amount change is larger, the weight will also be larger, which can better reflect the impact brought by the amount change during event switching. For example, when switching from a deposit event to a loan event and the amount changes significantly, this calculation method can make the weight more clearly reflect this change; when the event type remains unchanged: adjust based on the change rate of the amount of the adjacent timestamp on the basis of the original single timestamp amount. If the amount increases, the weight will increase accordingly, reflecting the enhanced capital investment of the user in the same event type; if the amount decreases, the weight will be appropriately reduced. Specifically, It reflects the change rate of the amounts at adjacent timestamps. Adding 1 to it means that, with reference to the amount at the (t - 1)-th timestamp, the weight is adjusted based on the change in the amount on the basis of the original amount at the t-th timestamp. For example, if this change rate is positive, it indicates that the amount has increased. It will be greater than 1, causing the weight to further increase under the influence of the original amount, reflecting the enhancement of the user's capital investment in the same type of event. This helps to more precisely depict the changes in the user's capital operations in the same type of financial behavior.

[0020] It should be further noted that in the present invention, there are only two types of events, namely deposits and loans, and there is only one type of event at each timestamp. Therefore, what needs to be concerned about is not only the event type itself (deposit or loan), but also the "behavior pattern" and "impact degree" behind the event. Specifically: The switching of event types (such as from deposit to loan) represents a "change in user behavior", and this change may mean a change in risk, demand, or strategy. The size of the event amount reflects the "intensity" or "importance" of the behavior change. A small amount may be a regular behavior, while a large amount may be an abnormal behavior (such as a sudden large loan). For example: When the event switches from a deposit to a loan and it is a large loan (such as switching from a 1000-yuan deposit to a 500,000-yuan loan), it may indicate that the user's funds are tight and the risk has increased; when the event switches from a loan to a deposit and it is a large deposit (such as immediately depositing 1 million after repaying the loan), it may reflect the user's capital inflow and the risk has decreased.

[0021] As a preferred solution of the method for model-based data service management for the digital intelligence financial ecosystem described in the present invention, based on the event switching weight W(t - 1→t) from the (t - 1)-th timestamp to the t-th timestamp, calculate the weight fluctuation range between adjacent event switching weights of the user. The calculation formula is as follows:

[0022] ΔW(t - 1→t + 1) = |W(t - 1→t) - W(t→t + 1)|;

[0023] Among them, ΔW(t - 1→t + 1) represents the weight fluctuation range between adjacent event switching weights, and W(t→t + 1) represents the event switching weight from the t-th timestamp to the (t + 1)-th timestamp.

[0024] Based on the weight fluctuation range ΔW(t - 1→t + 1) between adjacent event switching weights, calculate the cumulative weight fluctuation range of the user at all timestamps. The calculation formula is:

[0025] Based on the cumulative weight fluctuation range ΔAW(T) of the user at all timestamps, predict the event switching trend of the user at the (T + 1)-th timestamp. The calculation formula is as follows:

[0026]

[0027] Among them, P(T + 1) represents the event switching trend of the user at the (T + 1)-th timestamp. represents a preset adjustment coefficient, λ represents a preset exponential decay coefficient, μ represents a preset amount impact factor, and M(FSE T,i ) is the event amount of the financial service event corresponding to the T-th timestamp.

[0028] It should be noted that is used to adjust the influence degree of this part on the final predicted value P(T + 1), and e -λ(T-t) represents the exponential decay function, where (T - t) represents the time interval between timestamp T and timestamp t. As the time interval increases, the value of e -λ(T-t) decreases, meaning that the historical data farther from the current time contributes less to the prediction, highlighting the importance of recent data. ΔAW(T) reflects the fluctuation of the user's financial behavior in the past period; the greater the fluctuation amplitude, the higher the uncertainty of the user's financial behavior may be. μ is used to adjust the influence weight of the financial service event amount M(FSE T,i ) at the current timestamp T on the predicted value; emphasizes the influence of the user's financial behavior fluctuation at recent historical timestamps on the prediction. In the financial scenario, the user's behavior pattern may change rapidly, and recent behavior can better reflect future trends. This design makes the prediction closer to reality; combining the historical behavior fluctuation situation and the current event amount M(FSE T,i ), the former reflects the dynamic change characteristics of the user's behavior, and the latter shows the current fund operation level, multi-dimensionally depicting the user's financial behavior and providing a comprehensive basis for prediction.

[0029] Preset the cumulative weight fluctuation amplitude threshold and the event switching trend threshold. If the event switching trend P(T + 1) of the user at the (T + 1)-th timestamp is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation amplitude ΔAW(T) of the user at all timestamps is greater than or equal to the cumulative weight fluctuation amplitude threshold, then it is determined that there is a high-risk event switching for the user at the (T + 1)-th timestamp.

[0030] Obtain all users with high-risk event switching at the (T + 1)-th timestamp and conduct unified management.

[0031] A model-based data service management system for the digital intelligence financial ecosystem. This system includes: a data acquisition module, an event chain and matrix construction module, an event switching weight calculation module, and an event switching trend calculation and analysis management module;

[0032] The data acquisition module: acquires financial service events generated by the user on the digital intelligence financial ecological platform and the timestamps when the financial service events are generated, where the financial service events include the user's deposit events and loan events.

[0033] The event chain and matrix construction module: constructs a financial service event chain corresponding to the user at a single timestamp, where the financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events; constructs a financial service event chain matrix.

[0034] The event switching weight calculation module: calculates the event switching weights of the user at adjacent timestamps based on the financial service event chain matrix.

[0035] The event switching trend calculation and analysis management module: calculates the weight fluctuation amplitude between adjacent event switching weights of the user based on the event switching weights; calculates the cumulative weight fluctuation amplitude of the user at all timestamps based on the weight fluctuation amplitude; predicts the future event switching trend of the user based on the cumulative weight fluctuation amplitude; presets a threshold value for analysis and unified management.

[0036] Further, the data acquisition module includes a data acquisition unit.

[0037] The data acquisition unit: based on the digital intelligence financial ecological platform, acquires the user's behavior events, where the behavior events include the financial service events generated by the user on the digital intelligence financial ecological platform and the timestamps when the financial service events are generated, and the financial service events include the user's deposit events and loan events.

[0038] Further, the event chain and matrix construction module includes an event chain construction unit and a matrix construction unit.

[0039] The event chain construction unit: constructs a financial service event chain corresponding to the user at the t-th timestamp based on the timestamp; the financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events.

[0040] The matrix construction unit: acquires the financial service event chains corresponding to all timestamps, sorts the timestamps according to the ascending time rule, constructs a financial service event chain matrix, and uses the financial service event chain as the row of the matrix.

[0041] Further, the event switching weight calculation module includes an event switching weight calculation unit.

[0042] The event switching weight calculation unit: The event switching weight calculation formula includes the event switching weight calculation formula when the financial service events are the same under adjacent timestamps and the event switching weight calculation formula when the financial service events are different under adjacent timestamps.

[0043] Further, the event switching trend calculation and analysis management module includes an event switching trend calculation unit and an analysis management unit.

[0044] The event switching trend calculation unit: Based on the event switching weight, calculate the weight fluctuation amplitude between adjacent event switching weights of the user, and calculate the cumulative weight fluctuation amplitude of the user under all timestamps; Based on the cumulative weight fluctuation amplitude of the user under all timestamps, predict the future event switching trend of the user.

[0045] The analysis management unit: Preset the cumulative weight fluctuation amplitude threshold and the event switching trend threshold. If the future event switching trend of the user is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation amplitude of the user under all timestamps is greater than or equal to the cumulative weight fluctuation amplitude threshold, then it is determined that the user has a high-risk event switching in the future; Obtain all users with high-risk event switching in the future and conduct unified management.

[0046] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: In a model-based data service management method and system for the digital intelligence financial ecosystem provided by the present invention, the financial service events and timestamps of users on the digital intelligence financial ecosystem platform are obtained, providing basic data for subsequent analysis; Construct the financial service event chain of the user under a single timestamp, and further form a time-series financial service event chain matrix, thereby systematically recording the financial behavior patterns of the user at different times and laying a structured foundation for dynamic analysis; Based on the financial service event chain matrix, calculate the event switching weight of the user under adjacent timestamps, considering the dual influences of event type switching and amount change, and accurately measure the change trend of the user's financial behavior. For example, when a user switches from a deposit to a large loan, the system can identify the potential risk of tight funds, while switching from a loan to a large deposit may indicate the return of funds and reduce the risk; Further, based on the event switching weight, calculate the weight fluctuation amplitude and the cumulative weight fluctuation amplitude, and combine the exponential decay function to predict the future financial event switching trend, ensuring that the analysis is more in line with the actual change characteristics of the user's behavior. Through this method, the present invention can accurately warn of potential financial risks, optimize the risk control and service strategies under the digital intelligence financial ecosystem, and thus improve the intelligent level of financial management. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention.

[0048] Figure 1 It is a schematic diagram of the steps of a method for model-based data service management for the digital intelligence financial ecosystem of the present invention;

[0049] Figure 2 It is a schematic diagram of the structure of a model-based data service management system for the digital intelligence financial ecosystem of the present invention. Detailed implementation manners

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0051] Please refer to Figure 1 , in the first embodiment: A method for model-based data service management for the digital intelligence financial ecosystem is provided, and the method includes the following steps:

[0052] Step S1: Obtain the financial service events generated by the user on the digital intelligence financial ecosystem platform and the time stamps when the financial service events are generated, where the financial service events include the user's deposit events and loan events.

[0053] Specifically, based on the digital intelligence financial ecosystem platform, obtain the user's behavior events, where the behavior events include the financial service events generated by the user on the digital intelligence financial ecosystem platform and the time stamps when the financial service events are generated, and the financial service events include the user's deposit events and loan events.

[0054] Further, record the financial service event corresponding to the t-th time stamp as FSE t,i , where t ∈ [1, T], i ∈ [1, 2], T represents the total number of time stamps, i represents the index variable for distinguishing deposit events and loan events, and FSE t,1 indicates that the financial service event corresponding to the t-th time stamp is a deposit event, and FSE t,2 indicates that the financial service event corresponding to the t-th time stamp is a loan event.

[0055] Step S2: Construct a financial service event chain corresponding to the user at a single time stamp, where the financial service event chain includes a single time stamp, the financial service event corresponding to the single time stamp, and the event amount of the financial service event; construct a financial service event chain matrix.

[0056] Specifically, based on the time stamp, construct a financial service event chain corresponding to the user at the t-th time stamp, denoted as {(TI t , FSEt,i , M(FSE t,i )) | i ∈ [1, 2]}; Obtain the financial service event chains corresponding to all T timestamps, sort the timestamps according to the ascending time rule, construct a financial service event chain matrix, and use the financial service event chain as the row of the matrix, specifically as follows:

[0057]

[0058] Among them, TI t represents the t-th timestamp, and M(FSE t,i ) represents the event amount of the financial service event corresponding to the t-th timestamp.

[0059] Step S3: Based on the financial service event chain matrix, calculate the event switching weight of the user at adjacent timestamps.

[0060] Specifically, based on the financial service event chain matrix, calculate the event switching weight of the user at adjacent timestamps. The calculation formula of the event switching weight is as follows:

[0061] If the financial service events at adjacent timestamps are the same, that is, FSE t,i = FSE t-1,i , then the event switching weight formula is as follows:

[0062]

[0063] Among them, W(t - 1 → t) represents the event switching weight from the (t - 1)-th timestamp to the t-th timestamp, α represents the weight coefficient when the event type remains unchanged, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t - 1)-th timestamp.

[0064] If the financial service events at adjacent timestamps are different, that is, FSE t,i ≠ FSE t-1,i , then the event switching weight formula is as follows:

[0065]

[0066] Among them, W(t - 1 → t) represents the event switching weight from the (t - 1)-th timestamp to the t-th timestamp, β represents the weight coefficient when the event type changes, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t - 1)-th timestamp.

[0067] In the present invention, the formula represents that when the event type switches: take the average of the amounts of two adjacent timestamp events, and then multiply it by the absolute value of the relative amplitude of the amount change. This not only takes into account the size of the amount, but also highlights the degree of the amount change. When the amount change is larger, the weight will also be larger, which can better reflect the impact brought by the amount change during the event switch. For example, when switching from a deposit event to a loan event and there is a large change in the amount, this calculation method can make the weight more significantly reflect this change; when the event type remains unchanged: on the basis of the original amount of a single timestamp, adjust it in combination with the change rate of the amounts of adjacent timestamps. If the amount increases, the weight will increase accordingly, reflecting the enhancement of the user's capital investment in the same event type; if the amount decreases, the weight will be appropriately reduced. Specifically, It reflects the change rate of the amounts of adjacent timestamps. Adding 1 to it is to adjust the weight based on the amount at the (t - 1)th timestamp as a reference and according to the amount change at the tth timestamp. For example, if this change rate is positive, it means the amount has increased, it will be greater than 1, causing the weight to further increase on the basis of the influence of the original amount, reflecting the enhancement of the user's capital investment in the same type of event. This helps to more precisely depict the change in the user's capital operation in the same type of financial behavior.

[0068] It should be further noted that in the present invention, there are only two types of events, namely deposit and loan, and there is only one event at each timestamp. Therefore, not only the event type itself (deposit or loan) needs to be concerned, but also the "behavior pattern" and "influence degree" behind the event. Specifically: The event type switch (such as from deposit to loan) represents "the change in the user's behavior". This change may mean risk, demand change or strategy adjustment, and the size of the event amount reflects the "intensity" or "importance" of the behavior change. A small amount may be a regular behavior, while a large amount may be an abnormal behavior (such as a sudden large loan); for example: when the event switches from deposit to loan and it is a large loan (such as switching from a 1000 - yuan deposit to a 500,000 - yuan loan), it may indicate that the user's funds are tight and the risk increases; when the event switches from loan to deposit and it is a large deposit (such as immediately depositing 1 million after repaying the loan), it may reflect the user's capital return and the risk decreases.

[0069] Step S4: Based on the event - switching weights, calculate the weight fluctuation amplitude between adjacent event - switching weights of the user; based on the weight fluctuation amplitude, calculate the cumulative weight fluctuation amplitude of the user under all timestamps; based on the cumulative weight fluctuation amplitude, predict the future event - switching trend of the user; preset a threshold value, and analyze and conduct unified management.

[0070] Specifically, based on the event switching weight W(t - 1 → t) from the (t - 1)-th timestamp to the t-th timestamp, calculate the weight fluctuation amplitude between adjacent event switching weights of the user. The calculation formula is as follows:

[0071] ΔW(t - 1 → t + 1) = |W(t - 1 → t) - W(t → t + 1)|;

[0072] Among them, ΔW(t - 1 → t + 1) represents the weight fluctuation amplitude between adjacent event switching weights, and W(t → t + 1) represents the event switching weight from the t-th timestamp to the (t + 1)-th timestamp.

[0073] Furthermore, based on the weight fluctuation amplitude ΔW(t - 1 → t + 1) between adjacent event switching weights, calculate the cumulative weight fluctuation amplitude of the user under all timestamps. The calculation formula is:

[0074] Based on the cumulative weight fluctuation amplitude ΔAW(T) of the user under all timestamps, predict the event switching trend of the user at the (T + 1)-th timestamp. The calculation formula is as follows:

[0075]

[0076] Among them, P(T + 1) represents the event switching trend of the user at the (T + 1)-th timestamp, represents a preset adjustment coefficient, λ represents a preset exponential decay coefficient, μ represents a preset amount impact factor, M(FSE T,i ) is the event amount of the financial service event corresponding to the T-th timestamp.

[0077] It should be noted that, used to adjust the influence degree of this part on the final predicted value P(T + 1), e -λ(T-t) represents the exponential decay function, where (T - t) represents the time interval between timestamp T and timestamp t. As the time interval increases, e -λ(T-t) value decreases, meaning that the historical data farther from the current time contributes less to the prediction, highlighting the importance of recent data. ΔAW(T) reflects the fluctuation of the user's financial behavior in the past period; the greater the fluctuation amplitude, the higher the uncertainty of the user's financial behavior may be. μ is used to adjust the influence weight of the financial service event amount M(FSE T,i ) of the current timestamp T on the predicted value; emphasizes the influence of the user's financial behavior fluctuation under recent historical timestamps on the prediction. In the financial scenario, the user's behavior pattern may change rapidly, and recent behavior can better reflect future trends. This design makes the prediction closer to reality; taking into account the historical behavior fluctuation situation combined with the current event amount M(FSE T,i ), the former reflects the dynamic change characteristics of user behavior, and the latter shows the current magnitude of fund operations, multi-dimensionally depicting the user's financial behavior and providing a comprehensive basis for prediction.

[0078] Furthermore, a preset cumulative weight fluctuation amplitude threshold and an event switching trend threshold are set. If the event switching trend P(T + 1) of the user at the (T + 1)-th timestamp is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation amplitude ΔAW(T) of the user at all timestamps is greater than or equal to the cumulative weight fluctuation amplitude threshold, it is determined that there is a high-risk event switch for the user at the (T + 1)-th timestamp.

[0079] Obtain all users who have a high-risk event switch at the (T + 1)-th timestamp and conduct unified management.

[0080] In the present invention, when the user is a high-risk user and the event switching trend P(T + 1) is greater than or equal to the event switching trend threshold, it generally indicates that the user has a greater possibility of switching financial service events at the (T + 1)-th timestamp, and this switch may be accompanied by higher risks, as follows:

[0081] High-risk users themselves have higher default risks, fraud risks or other financial risks. At this time, if the event switching trend P(T + 1) is high, it means that from the comprehensive situation of historical behavior fluctuations and current financial service events, the user's financial behavior is about to change significantly. Such changes may be changes in the user's fund flow, transaction frequency, transaction amount, etc., and these changes often further exacerbate the risks. For example, a user with a poor credit status originally, whose event switching trend P(T + 1) is high, may mean that he is about to conduct a large transaction or frequently conduct different types of financial transactions, which will expose financial institutions to higher default risks and fund loss risks; it may also be that the user may be preparing to switch from one financial service to another more risky service, or may be covering up certain improper behaviors, such as money laundering and other illegal activities, by frequently switching services. For financial institutions, this requires high vigilance and timely measures for risk prevention and monitoring.

[0082] This situation warns financial institutions to strengthen risk management and monitoring measures for this user. Financial institutions may need to further investigate the source and use of the user's funds, evaluate their repayment ability and credit status, closely monitor their account activities, and prevent possible risk events such as default and fraud. At the same time, they may also need to re-examine the financial service strategy for this user, such as whether to adjust the credit limit, increase the loan interest rate or restrict certain high-risk financial services.

[0083] Please refer toFigure 2 In the second embodiment: A model-based data service management system for the digital intelligence financial ecosystem is provided. The system includes: a data acquisition module, an event chain and matrix construction module, an event switching weight calculation module, and an event switching trend calculation, analysis and management module.

[0084] The data acquisition module: acquires financial service events generated by a user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the user's deposit events and loan events.

[0085] The event chain and matrix construction module: constructs a financial service event chain corresponding to the user at a single timestamp. The financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events; constructs a financial service event chain matrix.

[0086] The event switching weight calculation module: calculates the event switching weights of the user at adjacent timestamps based on the financial service event chain matrix.

[0087] The event switching trend calculation, analysis and management module: calculates the weight fluctuation range between adjacent event switching weights of the user based on the event switching weights; calculates the cumulative weight fluctuation range of the user at all timestamps based on the weight fluctuation range; predicts the future event switching trend of the user based on the cumulative weight fluctuation range; presets a threshold value for analysis and unified management.

[0088] Further, the data acquisition module includes a data acquisition unit.

[0089] The data acquisition unit: based on the digital intelligence financial ecosystem platform, acquires the user's behavior events. The behavior events include the financial service events generated by the user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the user's deposit events and loan events.

[0090] Further, the event chain and matrix construction module includes an event chain construction unit and a matrix construction unit.

[0091] The event chain construction unit: based on the timestamp, constructs a financial service event chain corresponding to the user at the t-th timestamp. The financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events.

[0092] The matrix construction unit: acquires the financial service event chains corresponding to all timestamps, sorts the timestamps according to the ascending time rule, constructs a financial service event chain matrix, and uses the financial service event chain as the row of the matrix.

[0093] Further, the event switching weight calculation module includes an event switching weight calculation unit.

[0094] The event switching weight calculation unit: The event switching weight calculation formula includes the event switching weight calculation formula when the financial service events are the same under adjacent timestamps and the event switching weight calculation formula when the financial service events are different under adjacent timestamps.

[0095] Further, the event switching trend calculation and analysis management module includes an event switching trend calculation unit and an analysis management unit.

[0096] The event switching trend calculation unit: Based on the event switching weight, calculate the weight fluctuation range between adjacent event switching weights of the user, and calculate the cumulative weight fluctuation range of the user under all timestamps; Based on the cumulative weight fluctuation range of the user under all timestamps, predict the future event switching trend of the user.

[0097] The analysis management unit: Preset a cumulative weight fluctuation range threshold and an event switching trend threshold. If the future event switching trend of the user is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation range of the user under all timestamps is greater than or equal to the cumulative weight fluctuation range threshold, then it is determined that there is a high-risk event switching for the user in the future; Obtain all users with high-risk event switching in the future and conduct unified management.

[0098] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprises", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0099] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for managing modeled data services in the digital intelligence financial ecosystem, characterized in that, The method includes the following steps: Step S1: Obtain the financial service events generated by the user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the user's deposit events and loan events; Step S2: Construct a financial service event chain corresponding to the user at a single timestamp. The financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events; construct a financial service event chain matrix; Step S3: Calculate the event switching weights of the user at adjacent timestamps based on the financial service event chain matrix; Step S4: Calculate the weight fluctuation amplitude between adjacent event switching weights based on the event switching weights; calculate the cumulative weight fluctuation amplitude of the user at all timestamps based on the weight fluctuation amplitude; predict the future event switching trend of the user based on the cumulative weight fluctuation amplitude; preset a threshold, and analyze and conduct unified management.

2. The modeling data service management method for the digital intelligence financial ecosystem according to claim 1, wherein The specific implementation process of step S1 includes: Based on the digital intelligence financial ecosystem platform, obtain the user's behavior events. The behavior events include the financial service events generated by the user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated. The financial service events include the user's deposit events and loan events; Denote the financial service event corresponding to the t-th timestamp as FSE t,i , where t ∈ [1, T], i ∈ [1, 2], T represents the total number of timestamps, i represents the index variable for differentiating deposit events and loan events, and FSE t,1 indicates that the financial service event corresponding to the t-th timestamp is a deposit event, and FSE t,2 indicates that the financial service event corresponding to the t-th timestamp is a loan event.

3. A method for managing modeled data services for the digital intelligence financial ecosystem according to claim 2, characterized in that, The specific implementation process of step S2 includes: Based on the time stamp, construct a financial service event chain corresponding to the user at the t-th time stamp, denoted as {(TI t ,FSE t,i ,M(FSE t,i ))|i∈[1,2]}; Obtain the financial service event chains corresponding to all T time stamps, sort the time stamps according to the ascending time rule, construct a financial service event chain matrix, and use the financial service event chain as the row of the matrix, specifically as follows: Among them, TI t represents the t-th timestamp, and M(FSE t,i ) represents the event amount of the financial service event corresponding to the t-th timestamp.

4. A method for model-based data service management for the digital intelligence financial ecosystem according to claim 3, characterized in that, The specific implementation process of step S3 includes: Calculate the event switching weights of the user at adjacent timestamps based on the financial service event chain matrix. The calculation formula for the event switching weights is as follows: If the financial service events at adjacent timestamps are the same, i.e., FSE t,i = FSE t-1,i , then the event switching weight formula is as follows: Among them, W(t - 1→t) represents the event switching weight from the (t - 1)-th timestamp to the t-th timestamp, α represents the weight coefficient when the preset event type remains unchanged, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t - 1)-th timestamp; If the financial service events at adjacent timestamps are different, i.e., FSE t,i ≠FSE t-1,i , then the event switching weight formula is as follows: Among them, W(t-1→t) represents the event switching weight from the (t-1)-th timestamp to the t-th timestamp, β represents the weight coefficient when the preset event type changes, and M(FSE t-1,i ) represents the event amount of the financial service event corresponding to the (t-1)-th timestamp.

5. A method for modeling data service management in a digital intelligence financial ecosystem according to claim 4, characterized in that The specific implementation process of step S4 includes: Calculate the weight fluctuation amplitude between adjacent event switching weights based on the event switching weight W(t - 1→t) from the (t - 1)-th timestamp to the t-th timestamp. The calculation formula is as follows: ΔW(t - 1→t + 1) = |W(t - 1→t) - W(t→t + 1)|; Where, ΔW(t - 1→t + 1) represents the weight fluctuation amplitude between adjacent event switching weights, and W(t→t + 1) represents the event switching weight from the t-th timestamp to the (t + 1)-th timestamp; Based on the weight fluctuation amplitude ΔW(t-1→t+1) between adjacent event switching weights, calculate the cumulative weight fluctuation amplitude of the user at all timestamps. The calculation formula is as follows: Predict the event switching trend of the user at the (T + 1)-th timestamp based on the cumulative weight fluctuation amplitude ΔAW(T) of the user at all timestamps. The calculation formula is as follows: Among them, P(T + 1) represents the event switching trend of the user at the (T + 1)-th timestamp. represents a preset adjustment coefficient, λ represents a preset exponential decay coefficient, μ represents a preset amount impact factor, and M(FSE T,i ) is the event amount of the financial service event corresponding to the T-th timestamp. Preset a cumulative weight fluctuation amplitude threshold and an event switching trend threshold. If the event switching trend P(T + 1) of the user at the (T + 1)-th timestamp is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation amplitude ΔAW(T) of the user at all timestamps is greater than or equal to the cumulative weight fluctuation amplitude threshold, then it is determined that there is a high-risk event switching for the user at the (T + 1)-th timestamp; Obtain all users with high-risk event switching at the (T + 1)-th timestamp and conduct unified management.

6. A model-based data service management system for the digital intelligence financial ecosystem, which executes a model-based data service management method for the digital intelligence financial ecosystem according to any one of claims 1-5, characterized in that, The system includes: a data acquisition module, an event chain and matrix construction module, an event switching weight calculation module, and an event switching trend calculation and analysis management module; The data acquisition module: acquires financial service events generated by a user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated, where the financial service events include the user's deposit events and loan events; The event chain and matrix construction module: constructs a financial service event chain corresponding to the user at a single timestamp, where the financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events; constructs a financial service event chain matrix; The event switching weight calculation module: calculates the event switching weights of the user at adjacent timestamps based on the financial service event chain matrix; The event switching trend calculation, analysis and management module: calculates the weight fluctuation range between adjacent event switching weights of the user based on the event switching weights; calculates the cumulative weight fluctuation range of the user at all timestamps based on the weight fluctuation range; predicts the future event switching trend of the user based on the cumulative weight fluctuation range; preset thresholds for analysis and unified management.

7. A model-based data service management system for the digital intelligence financial ecosystem according to claim 6, characterized in that: The data acquisition module includes a data acquisition unit; The data acquisition unit: based on the digital intelligence financial ecosystem platform, acquires the user's behavior events, where the behavior events include the financial service events generated by the user on the digital intelligence financial ecosystem platform and the timestamps when the financial service events are generated, and the financial service events include the user's deposit events and loan events.

8. A model-based data service management system for the digital intelligence financial ecosystem according to claim 7, characterized in that: The event chain and matrix construction module includes an event chain construction unit and a matrix construction unit; The event chain construction unit: constructs a financial service event chain corresponding to the user at the t-th timestamp based on the timestamp; The financial service event chain includes a single timestamp, the financial service events corresponding to the single timestamp, and the event amounts of the financial service events; The matrix construction unit: acquires the financial service event chains corresponding to all timestamps, sorts the timestamps according to the ascending time rule, constructs a financial service event chain matrix, and uses the financial service event chain as the row of the matrix.

9. A model-based data service management system for the digital intelligence financial ecosystem according to claim 8, characterized in that: The event switching weight calculation module includes an event switching weight calculation unit; The event switching weight calculation unit: the event switching weight calculation formula includes the event switching weight calculation formula when the financial service events at adjacent timestamps are the same and the event switching weight calculation formula when the financial service events at adjacent timestamps are different.

10. A model-based data service management system for the digital intelligence financial ecosystem according to claim 9, characterized in that: The event switching trend calculation, analysis and management module includes an event switching trend calculation unit and an analysis and management unit; The event switching trend calculation unit: calculates the weight fluctuation range between adjacent event switching weights of the user based on the event switching weights, and calculates the cumulative weight fluctuation range of the user at all timestamps; Predicts the future event switching trend of the user based on the cumulative weight fluctuation range of the user at all timestamps; The analysis and management unit: preset a cumulative weight fluctuation range threshold and an event switching trend threshold. If the future event switching trend of the user is greater than or equal to the event switching trend threshold, and the cumulative weight fluctuation range of the user at all timestamps is greater than or equal to the cumulative weight fluctuation range threshold, it is determined that there is a high-risk event switching for the user in the future; Obtain all users who will experience high-risk event switches in the future and conduct unified management.

Images