Scoring processing method and device, storage medium and electronic equipment

Abstract

The invention discloses a scoring processing method and device, a storage medium and electronic equipment, and the method comprises the steps: determining basic user credit scoring data for a target user based on a credit scoring model, determining a user credit feature for the target user, and determining a credit score of the target user based on the user credit feature; and determining score change factor data for the basic user credit score data by adopting a score change factor model, and performing score adjustment by adopting a credit score correction model based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain a target user credit score for the target user.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a scoring processing method, device, storage medium and electronic device. Background Art

[0002] Risk control (referred to as risk control) on platform services such as consumer finance services, online shopping services, and express logistics services mainly relies on complex, massive, high-dimensional data information to model the credit risk of users, and is implemented based on the credit scoring model developed by the modeling. The credit scoring model can be used to perform credit scoring, which represents the credit value or credit risk of an individual or group, and has great reference value in trust assessment. For example, enterprises and financial institutions use this credit score to decide whether to provide or grant credit to a certain person or group. Summary of the invention

[0003] This specification provides a scoring processing method, device, storage medium and electronic device, and the technical solution is as follows:

[0004] In a first aspect, this specification provides a scoring processing method, the method comprising:

[0005] Determine basic user credit score data for the target user based on the credit score model, and determine user credit characteristics for the target user;

[0006] Based on the user's credit characteristics, a score change factor model is used to determine score change factor data for the basic user's credit score data, wherein the score change factor data is confidence level data that the score change factor model identifies the change in the basic user's credit score to represent the change in the user's real credit risk;

[0007] Based on the basic user credit score data, the score change factor data and the historical user credit score data, a credit score correction model is used to perform score adjustment to obtain a target user credit score for the target user. The historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

[0008] In a second aspect, this specification provides a scoring processing device, the device comprising:

[0009] A basic scoring module, used to determine basic user credit score data for a target user based on a credit scoring model, and determine user credit characteristics for the target user;

[0010] A factor determination module, for determining score change factor data for the basic user credit score data using a score change factor model based on user credit characteristics, wherein the score change factor data is confidence level data that the score change factor model identifies a change in the basic user credit score to represent a change in the user's real credit risk;

[0011] A score correction module is used to use a credit score correction model to perform score adjustment based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain a target user credit score for the target user, wherein the historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

[0012] In a third aspect, the present specification provides a computer storage medium, wherein the computer storage medium stores at least one instruction, wherein the instruction is suitable for being loaded by a processor and executing the method steps of one or more embodiments of the present specification.

[0013] In a fourth aspect, the present specification provides a computer program product, wherein the computer program product stores at least one instruction, wherein the instruction is suitable for being loaded by a processor and executing the method steps of one or more embodiments of the present specification.

[0014] In a fifth aspect, the present specification provides an electronic device, which may include: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the method steps of one or more embodiments of the present specification.

[0015] The beneficial effects brought by the technical solutions provided by some embodiments of this specification include at least:

[0016] In one or more embodiments of the present specification, the service platform first determines the basic user credit score data for the target user based on the credit score model, determines the score change factor data for the basic user credit score data based on the credit characteristics of the target user using the score change factor model, and then uses the credit score correction model to adjust the score based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain the target user credit score for the target user, thereby overcoming the disadvantage of the traditional credit score model in that the fluctuation of the underlying data leads to the fluctuation of the credit score of the individual user, and introducing the score change factor model to identify the score change factor corresponding to the actual risk change of the change in the user's credit score, and adjusting the basic user credit score according to the actual credit risk change of the user to obtain an accurate target user credit score, thereby ensuring the stability and rationality of the user's credit score, so that the user's credit score is more in line with the actual risk change, and can provide a more reliable basis for credit decision-making and risk management. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in this specification or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a scenario diagram of a scoring processing system provided in this specification;

[0019] Figure 2 It is a flowchart of a scoring processing method provided in this manual;

[0020] Figure 3 This is a schematic diagram of a process for determining the score change factor data provided in this manual;

[0021] Figure 4 This is a flow chart of determining the characteristics of a user's overdue status change provided in this specification;

[0022] Figure 5 It is a schematic diagram of a scenario processed by a fractional change factor model provided in this specification;

[0023] Figure 6 It is a flowchart of a credit score correction model used in this specification;

[0024] Figure 7 It is a flowchart of a score change factor model training provided in this specification;

[0025] Figure 8 It is a structural schematic diagram of a scoring processing device provided in this specification;

[0026] Fig. 9 It is a structural schematic diagram of an electronic device provided in this manual. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in this specification to clearly and completely describe the technical solutions in this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0028] In the description of this specification, it should be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In the description of this specification, it should be noted that, unless otherwise clearly specified and limited, "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices. For those of ordinary skill in the art, the specific meanings of the above terms in this specification can be understood in specific circumstances. In addition, in the description of this specification, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are an "or" relationship.

[0029] In related technologies, due to the complexity of user data, the rapid changes in behavioral data and the influence of the external environment, the credit scoring model constructed using these data faces major challenges. Specifically, even if the user's actual credit risk has not changed significantly, the user's credit score will still fluctuate dramatically due to changes in the underlying data characteristics of the entire platform's overall transactions (which can be considered a change in the overall transaction data characteristics rather than a change in the user's personal data characteristics). This volatility greatly reduces the stability, rationality and accuracy of the credit score.

[0030] The present specification is described in detail below with reference to specific embodiments.

[0031] See also Figure 1 , is a schematic diagram of a scoring processing system provided in this specification. Figure 1 As shown, the scoring processing system may include at least a client cluster and a service platform 100 .

[0032] The client cluster may include at least one client, such as Figure 1 As shown, it specifically includes client 1 corresponding to user 1, client 2 corresponding to user 2, ..., client n corresponding to user n, where n is an integer greater than 0.

[0033] Each client in the client cluster may be an electronic device with communication function, including but not limited to: wearable device, handheld device, personal computer, tablet computer, vehicle-mounted device, smart phone, computing device or other processing device connected to wireless modem, etc. Electronic devices may be called different names in different networks, such as: user equipment, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, electronic device, wireless communication device, user agent or user device, cellular phone, cordless phone, personal digital assistant (PDA), electronic device in 5G network or future evolution network, etc.

[0034] The service platform 100 can be a separate server device, such as a rack-mounted, blade, tower, or cabinet-mounted server device, or a workstation, mainframe computer, or other hardware device with strong computing power; it can also be a server cluster composed of multiple servers, and the servers in the service cluster can be composed in a symmetrical manner, wherein each server has equivalent functions and status in the transaction link, and each server can provide services to the outside independently, and the independent service can be understood as not requiring the assistance of other servers.

[0035] In one or more embodiments of the present specification, the service platform 100 may establish a communication connection with at least one client in the client cluster, and complete data interaction in the scoring process based on the communication connection, such as online transaction data interaction. For example, the service platform 100 may obtain user scoring for the user of the client based on the scoring processing method of the present specification.

[0036] It should be noted that the service platform 100 and at least one client in the client cluster establish a communication connection through a network for interactive communication, wherein the network can be a wireless network or a wired network, the wireless network includes but is not limited to a cellular network, a wireless local area network, an infrared network or a Bluetooth network, and the wired network includes but is not limited to Ethernet, a universal serial bus (USB) or a controller local area network. In one or more embodiments of the specification, technologies and / or formats including Hyper Text Mark-up Language (HTML), Extensible Markup Language (XML), etc. are used to represent data (such as a target compressed package) exchanged through a network. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above data communication technologies.

[0037] The scoring processing system embodiment provided in this specification and the scoring processing method in one or more embodiments belong to the same concept. The execution subject corresponding to the scoring processing method involved in one or more embodiments of the specification can be the above-mentioned service platform 100; the execution subject corresponding to the scoring processing method involved in one or more embodiments of the specification can also be the electronic device corresponding to the client, which is determined based on the actual application environment. The embodiment of the scoring processing system and its implementation process can be detailed in the following method embodiment, which will not be repeated here.

[0038] based on Figure 1 The scene diagram is shown, and the scoring processing method provided by one or more embodiments of this specification is introduced in detail below.

[0039] See also Figure 2 , is a flowchart of a scoring processing method for one or more embodiments of the present specification, which can be implemented by a computer program and can be run on a scoring processing device based on the von Neumann system. The computer program can be integrated into an application or run as an independent tool application. The scoring processing device can be a service platform.

[0040] Specifically, the scoring processing method includes:

[0041] S102: Determine basic user credit score data for the target user based on the credit score model, and determine user credit features for the target user;

[0042] The credit scoring model is a credit scoring model that is used for risk control (referred to as risk control) of platform transaction services and uses complex, massive, high-dimensional data information to model the credit risk of users. The credit scoring model can score the credit of users in transaction services. The credit score represents the credit value or credit risk of an individual or group, and has great reference value in trust assessment. For example, enterprises and financial institutions use this credit score to decide whether to provide or grant credit to a certain person or group.

[0043] It should be noted that the credit scoring model involved in this specification can be developed and customized based on actual platform transaction services, and is not specifically limited here.

[0044] In daily business applications, credit scoring models can be used by financial institutions such as banks to score the credit of loan applicants, so as to predict the risk of loan applicants being overdue based on the score.

[0045] The basic user credit score data can be understood as a score sequence composed of user credit scores at different times over a period of time.

[0046] In actual applications, user transaction behavior data at different times within a period of time is input into a credit scoring model, and user credit scoring data can be output. The user credit scoring data includes user credit scores at different times. In this specification, the output of the credit scoring model is not used directly but is used as basic user credit scoring data. The basic user credit scoring data is further used to offset the fluctuations of the underlying data of the market, eliminate the impact of non-user personal components, and then adjust the score to obtain the target user credit score.

[0047] In principle, user transaction behavior data refers to all transaction and credit activity records of a user over a period of time, including but not limited to consumption, repayment, loan application, account changes, transaction information, etc.

[0048] Specifically, after determining basic user credit score data for the target user based on the credit score model, determining user credit features for the target user;

[0049] User credit characteristics are key indicators used to assess user credit risk, usually including user underlying attribute characteristics and user credit indicator trend characteristics;

[0050] The underlying attributes of users can be understood as the basic attributes of users, such as: age: the age range of users, which is usually related to credit risk; sexual characteristics: the sexual characteristics of users, which may affect consumption behavior and credit performance; occupation: the type of occupation of users, some occupations may be related to credit stability. Living status: such as renting or owning a house, which affects the financial stability of users; fixed asset value; intangible asset value; historical maximum overdue days; number of default executions, etc.

[0051] The trend characteristics of user credit indicators can be understood as the user credit indicators that are usually set. The trend characteristics of user credit indicators refer to the indicator change trend characteristics of the user's credit indicators within a period of time. For example, the trend characteristics of credit utilization rate change: the upward or downward trend of the user's credit utilization rate within a period of time reflects the credit management ability; the trend characteristics of repayment history change: such as the change in the on-time repayment rate, showing the improvement or deterioration of the user's credit management; the trend characteristics of new account opening: the number of newly opened credit accounts within a period of time, and the frequent opening of new accounts may indicate an increase in credit risk; the trend characteristics of overdue record change: the increase or decrease trend of overdue payment records directly affects the credit score;

[0052] S104: Based on the user credit characteristics, using the score change factor model to determine the score change factor data for the basic user credit score data, the score change factor data being the confidence level data that the score change factor model identifies the basic user credit score change to represent the user's real credit risk change;

[0053] The score change factor model is used to identify the confidence level of the change in the user's credit score corresponding to the real risk change. Exemplarily, the user's credit characteristics and the basic user credit score data are input into the score change factor model, and the score change factor model identifies the user's real risk change (information) from the user's credit characteristics, and can identify the basic user credit score change from the basic user credit score data (information), and then analyzes the correlation change degree between the basic user credit score data (information) and the user's real risk change (information) to obtain the score change factor data; the score change factor data is the confidence level data of the score change factor model identifying the basic user credit score change to represent the user's real credit risk change, which can be understood as the confidence level that the basic user credit score change represents or replaces the user's real credit risk change. Generally, the greater the positive correlation between the basic user credit score data (information) and the user's real risk change (information), the higher the score change factor data; conversely, generally, the smaller the positive correlation between the basic user credit score data (information) and the user's real risk change (information), the lower the score change factor data;

[0054] Usually, when a user's credit score changes due to feature noise fluctuations, but the user's real risk does not change significantly, the corresponding model-predicted score change factor data will be smaller, and the user's credit score needs to remain stable. Conversely, when the user's actual credit risk changes, the corresponding model-predicted score change factor data will be larger, indicating that the user's credit score should change with the change in the user's real risk.

[0055] S106: Based on the basic user credit score data, the score change factor data and the historical user credit score data, a credit score correction model is used to perform score adjustment to obtain a target user credit score for the target user, wherein the historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

[0056] The basic user credit score data is the basic user credit score of the credit score model at multiple times within a period of time, usually a basic user credit score sequence, which can provide the credit status of the target user for the current transaction credit score model to adjust the credit score.

[0057] The score change factor data is a score change factor sequence of the confidence data of the change in the basic user credit score that the score change factor model identifies to represent the change in the user's real credit risk. It identifies the confidence level that affects the change in the credit score and can reflect the change in the user's real credit risk. Each score change factor in the score change factor data can correspond to the basic user credit score in the basic user credit score data.

[0058] Historical user credit score data is the actual credit score after the target user has been adjusted in the past. Basic user credit score data, score change factor data and historical user credit score data provide the time context of the user's credit performance.

[0059] In an illustrative manner, the basic user credit score data, the score change factor data and the historical user credit score data are input into the credit score correction model. The credit score correction model comprehensively considers the various model input data input, adjusts the basic user credit score corresponding to the target user at the current target time (usually the most recent credit score time), and obtains the target user credit score after score adjustment to more accurately reflect its credit risk. The target user credit score is a new score obtained after model adjustment, which reflects the user's current credit status and risk level.

[0060] In one or more embodiments of the present specification, the service platform first determines the basic user credit score data for the target user based on the credit score model, determines the score change factor data for the basic user credit score data based on the credit characteristics of the target user using the score change factor model, and then uses the credit score correction model to adjust the score based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain the target user credit score for the target user, thereby overcoming the disadvantage of the traditional credit score model in that the fluctuation of the underlying data leads to the fluctuation of the credit score of the individual user, and introducing the score change factor model to identify the score change factor corresponding to the actual risk change of the change in the user's credit score, and adjusting the basic user credit score according to the actual credit risk change of the user to obtain an accurate target user credit score, thereby ensuring the stability and rationality of the user's credit score, so that the user's credit score is more in line with the actual risk change, and can provide a more reliable basis for credit decision-making and risk management.

[0061] See also Figure 3 , Figure 3 This is a flow chart of determining a score change factor data proposed in one or more embodiments of this specification. Specifically, the following method can be used to determine the score change factor data for the basic user credit score data based on the user credit characteristics using the score change factor model:

[0062] S202: Determine the user's underlying attribute characteristics and the user's credit index trend characteristics from the user's credit characteristics, and determine the user's overdue status change characteristics based on the user's credit overdue information in the user's credit characteristics;

[0063] User credit overdue information: refers to the situation where the user fails to fulfill the contract on time, including the number of overdue times, time, amount of overdue and other information.

[0064] Exemplarily, the user's overdue status change feature can be understood as a Y mark introduced based on the user's credit expectation information;

[0065] The introduction of the user's overdue status change characteristics can, to a certain extent, reflect the changes in the user's real credit risk over time, which helps the score change factor model to identify and determine the user's real credit risk changes, and then compare the basic user credit score changes to represent the confidence level of the user's real credit risk changes.

[0066] In a possible implementation, Figure 4 As shown, Figure 4 It is a flow chart of determining the characteristics of the user's overdue status change, as follows:

[0067] S3002: Determine a first credit overdue status at a first evaluation time and a second credit overdue status at a second evaluation time based on user credit overdue information in the user credit feature, wherein the first evaluation time and the second evaluation time are separated by a preset time interval;

[0068] The user overdue status change feature defines that the user overdue status change can be adjusted based on the actual transaction. For example, the preset time considers the user's future 1 month, 2 months, 3 months, or even more distant months;

[0069] S3004: if the first credit overdue status matches the second credit overdue status, generating a user overdue status change feature of a stable status type;

[0070] S3006: If the first credit overdue status does not match the second credit overdue status, a user overdue status change feature of a status change type is generated.

[0071] For example: the user overdue status change feature defines whether the overdue status of month M (whether overdue) and the overdue status of month M+i (whether overdue) have changed. If the status has changed, the label is a value of the status change type (such as 1), otherwise it is a value of the status stability type (such as 0); for example: the second evaluation time is December (M+i), the preset time i is 3, December is overdue, and the first evaluation time is September is not overdue, then the value of the September (M) Y label is 0; for another example, if the current month is November, November is overdue, and August is not overdue, then the value of the August Y label is 1;

[0072] S204: Determine basic user credit score change information and user real credit risk change information for the basic user credit score data using a score change factor model based on the user's underlying attribute characteristics, the user's credit indicator trend characteristics, and the user's overdue status change characteristics;

[0073] Indicatively, the user's underlying attribute characteristics, the user's credit index trend characteristics, and the user's overdue index characteristics are used as input data of the score change factor model to provide a comprehensive description of the user's credit status. The user's underlying attribute characteristics, the user's credit index trend characteristics, and the user's overdue index characteristics are used as inputs to the score change factor model. The score change factor model first identifies the basic user credit score change information and the user's real credit risk change information for the basic user credit score data;

[0074] The basic user credit score change information is the score change factor model's assessment of changes in the user's credit score, including the extent and reasons for the score increase or decrease.

[0075] The user's real credit risk change information reflects the user's real credit risk change, and is predicted based on a comprehensive analysis of the user's underlying attribute characteristics, credit indicator trend characteristics, and overdue indicator characteristics;

[0076] S206: Based on the basic user credit score change information and the user's real credit risk change information, a score change factor model is used to identify confidence data of the basic user credit score change that represents the user's real credit risk change, and the confidence data is used as score change factor data for the basic user credit score data.

[0077] For example, Figure 5 As shown, Figure 5 It is a schematic diagram of the scenario processed by the score change factor model. The score change factor model determines the user's underlying attribute characteristics (such as Figure 6 The underlying characteristics shown in Figure 1) and the trend characteristics of user credit indicators (such as Figure 6 The score change factor model determines the user overdue status change characteristics (such as Figure 6 Based on this, based on the user's underlying attribute characteristics, the user's credit indicator trend characteristics and the user's overdue status change characteristics, a score change factor model is used to determine the basic user credit score change information and the user's real credit risk change information for the basic user credit score data, and then a comparative analysis is performed based on the basic user credit score change information and the user's real credit risk change information, and the score change factor model is used to identify the confidence level data of the basic user credit score change that represents the user's real credit risk change, that is, to determine the representativeness of the basic user credit score change to the user's real credit risk change, and use the confidence level data as the score change factor data for the basic user credit score data.

[0078] In one or more embodiments of this specification, the user's underlying attribute characteristics, credit indicator trend characteristics, and overdue indicator characteristics are input into the score change factor model to dynamically identify and determine the degree of correlation between the user's credit score change and the actual credit risk change, and generate confidence data as the basis for score adjustment. This effectively improves the accuracy and stability of the credit score, makes the score more realistically reflect the user's credit risk changes, reduces the instability of the score caused by external environment or data fluctuations, and thus provides more reliable support for credit affairs decision-making and risk management, and reduces potential credit risks.

[0079] See also Figure 6 , Figure 6 It is a flowchart of the use of the credit score correction model proposed in one or more embodiments of this specification.

[0080] S4002: Determine a basic user credit score time series from the basic user credit score data, and determine a score change factor time series from the score change factor data;

[0081] The basic user credit score data includes the basic user credit scores of the credit score model at multiple times within a period of time, based on which a basic user credit score time series can be determined from the basic user credit score data, and the basic user credit score time series is a sequence composed of basic user credit scores at multiple different times;

[0082] The score change factor data includes the score change factor of the basic user credit score at each different time in the basic user credit score time series, based on which the score change factor time series can be determined from the score change factor data.

[0083] S4004: Input the basic user credit score time series and the score change factor time series into a credit score correction model, and use the credit score correction model to perform score weighting processing on the target basic user credit score using a score weight parameter and historical user credit score data to obtain a target user credit score for the target user, wherein the target basic user credit score is the basic user credit score corresponding to the target score time unit in the basic user credit score data.

[0084] Illustratively, the historical user credit score data is the actual credit score after score adjustment for the target user in the past, and the basic user credit score data, score change factor data and historical user credit score data provide the time context of the user's credit performance.

[0085] The credit score correction model combines the basic user credit score and the time series of the score change factor, and performs weight learning training in a similar way to the cyclic model training (non-model training process) to obtain high-quality scoring weight parameters to balance the performance and stability of the credit score correction model between the historical user credit score and the basic user credit score.

[0086] In a feasible implementation, the following method may be referred to:

[0087] A2: Input the basic user credit score time series and the score change factor time series into a credit score correction model, determine the score weight parameter between the basic user credit score item and the historical user credit score item through the credit score correction model, and obtain the historical user credit score data corresponding to the score weight parameter and the historical user credit score item;

[0088] In a feasible implementation manner, the basic user credit score time series and the score change factor time series are input into a credit score correction model, and the score weight parameter between the basic user credit score item and the historical user credit score item is determined by the credit score correction model, which can refer to the following method:

[0089] 1. Input the basic user credit score time series and the score change factor time series into the credit score correction model for at least one round of weight learning training;

[0090] 2. During the weight learning training process, determine the current scoring weight parameters and credit behavior labels, determine the predicted user credit score data using the target formula based on the basic user credit score time series, the scoring weight parameters and the historical user credit score data, calculate the weight learning loss based on the predicted user credit score data, the credit behavior labels, the score change factor time series and the historical user credit score data, and use the weight learning loss to adjust the scoring weight parameters until the weight learning training end conditions are met.

[0091] Indicatively, the basic user credit score time series and the score change factor time series correspond to multiple parameters. The basic user credit score time series and the score change factor time series are input into the credit score correction model to perform at least one round of weight learning training of the score weight parameter (which can be regarded as the class model weight parameter). The score weight parameter of the first round of weight learning training is an initial value. In each round of weight learning training after the first round:

[0092] 1) Forward propagation training process: determine the current scoring weight parameters and credit behavior labels, and use the target formula to determine the predicted user credit score data based on the basic user credit score time series, scoring weight parameters and historical user credit score data;

[0093] Furthermore, the scoring weight parameter of the first round of weight learning training is an initial value; the credit behavior label is whether the target user has performed a credit behavior in a certain time unit that can be determined from the user's credit characteristics. If a credit behavior occurs, the credit behavior label is the first value (such as 1), and if no credit behavior occurs, the credit behavior label is the second value (such as 0); such as whether the target user has performed a credit loan behavior in the corresponding time under a certain time unit t.

[0094] 2) Back propagation training process: Calculate the weight learning loss based on the predicted user credit score data, the credit behavior label, the score change factor time series and the historical user credit score data, and use the weight learning loss to adjust the score weight parameter until the weight learning training end condition is met;

[0095] Optionally, the weight learning loss calculated based on the predicted user credit score data, the credit behavior label, the score change factor time series, and the historical user credit score data may refer to the following method:

[0096] In one or more embodiments of the present specification, the weighted learning loss can be obtained by using a weighted learning loss calculation formula based on the predicted user credit score data, credit behavior labels, score change factor time series, and historical user credit score data;

[0097] The weight learning loss calculation formula satisfies the following formula:

[0098]

[0099] Wherein, L represents the weight learning loss, T represents the upper limit of the time unit of the time series, t represents the tth time unit, N represents the number of time units of the time series, BCE() represents the cross entropy operation processing, and R t represents the predicted user credit score corresponding to the tth time unit in the predicted user credit score data, and the y t represents the credit behavior label value corresponding to the t-th time unit in the credit behavior label, and the λ t represents the fractional change factor value corresponding to the tth time unit in the fractional change factor time series, and the R t-1 Represents the predicted user credit score corresponding to the t-1th time unit in the predicted user credit score data.

[0100] The above weighted learning loss calculation formula proposes a method similar to a cyclic training model to weight historical scores and basic score information, balancing the performance of the model and the stability of the score. This design uses time series data to enhance the credit score correction model's ability to adapt to changes in user credit scores over time.

[0101] Furthermore, in addition to the absorbed cross entropy term, the weighted learning loss calculation formula also incorporates the mean square error constraint term, and introduces the score change factor as an adjustment coefficient to coordinate model performance and stability. When the predicted value of the score change factor is larger, it means that the probability of the user's real risk change is higher, and the model pays more attention to performance. Conversely, when the predicted value of the score change factor is smaller, it means that the probability of the user's real risk change is lower, and the model pays more attention to stability. This method can adaptively learn and improve the stability and rationality of risk scoring.

[0102] Furthermore, the credit score correction model is created based on a machine learning model. Different from the model training in the related art, in this specification, each score adjustment process for different users can be regarded as a small model training process. The model training process is an improved weight learning training. The weight learning training is mainly to use weight learning loss to adjust the score weight parameters until the weight learning training end conditions are met, thereby applying new score weight parameters to perform score adjustment.

[0103] The machine learning models used in the credit score correction model include but are not limited to the fitting of one or more machine learning models such as Convolutional Neural Network (CNN) model, Deep Neural Network (DNN) model, Recurrent Neural Networks (RNN) model, embedding model, Gradient Boosting Decision Tree (GBDT) model, Logistic Regression (LR) model, etc.;

[0104] Optionally, the credit score correction model may adopt a Similar-RNN model in a recurrent neural network.

[0105] Optionally, the weight learning training end condition may include, for example, the weight learning loss value of the loss function is less than or equal to a preset loss function threshold, the number of iterations reaches a preset number threshold, etc. The specific weight learning training end condition can be determined based on actual conditions and is not specifically limited here.

[0106] In one or more embodiments of the present specification, a target user credit score is obtained by using a target formula based on the target basic user credit score, the historical user credit score data, and the score weight parameter;

[0107] The target formula satisfies the following formula:

[0108] R=w*y+(1-w)*r

[0109] Among them, R represents the target user credit score, w represents the score weight parameter, r represents the historical user credit score data, and y is the target basic user credit score.

[0110] The scoring weight parameter w represents the relative importance of the basic user credit scoring item and the historical user credit scoring item in the target user's credit scoring, ranging from 0 to 1.

[0111] R represents the target user’s credit score, which is the final score after adjustment and reflects the user’s current credit risk;

[0112] r represents the historical user credit score data. Specifically, r in each round is the historical user credit score corresponding to the target time unit t, representing the user's past credit performance;

[0113] y is the basic user credit score corresponding to the target time unit t; further, in actual applications, the basic user credit score at the most recent time t is adjusted using the above target company to obtain the target user credit score for the target user, which represents the user's current credit status.

[0114] It can be understood that through the target formula, the current basic credit score and the historical credit score data are weighted and combined to obtain the target user credit score R. The credit score is adjusted by weight: when w is larger, the current credit score (y) has a greater impact on the final result; when w is smaller, the historical credit score data (r) has a greater impact on the final result.

[0115] Optionally, the time unit length of time training such as the basic user credit score time series and the score change factor time series can be preset, such as 6 time units (such as 6 months), which is equivalent to the expert experience that the credit risk of the user will not change significantly within half a year. Therefore, in the model architecture, the basic model scores and score change factor scores of the past 6 months are used as input in a time series manner, and the above loss function is used as the optimization target. The weight learning is performed in a manner similar to a recurrent neural network to obtain the final target user credit score;

[0116] In one or more embodiments of this specification, the user's basic credit score is combined with historical credit score data, and a score weight parameter is introduced to construct a method for dynamically adjusting the credit score. Through the calculation of the target formula, it is ensured that the target user's credit score can not only reflect the user's current credit status, but also fully consider the historical credit performance, and ensure the accuracy and stability of the score. This method can effectively deal with score fluctuations caused by changes in the external environment or user behavior, provide a more reliable basis for credit decision-making and risk management, and enhance the ability to identify user credit risks.

[0117] See also Figure 7 , Figure 7 It is a flowchart of the score change factor model training proposed in one or more embodiments of this specification.

[0118] S5002: Create an initial score change factor model;

[0119] S5004: Obtain user credit samples and label the score change factor data labels of the user credit samples, wherein the user credit samples include sample user credit features and sample basic user credit score data;

[0120] S5006: Perform at least one round of model training on the initial score change factor model based on the user credit sample, predict the predicted score change factor data for the sample basic user credit score data through the initial score change factor model, adjust the model parameters of the initial score change factor model based on the predicted score change factor data and the sample basic user credit score data, until the model training is completed, and obtain the score change factor model.

[0121] The following is an explanation of the model training process of the initial score change factor model:

[0122] Model creation: creating an initial score change factor model based on the machine learning model, where the (initial) score change factor data is the confidence level data of the score change factor model identifying the change in the basic user's credit score to represent the change in the user's real credit risk;

[0123] Sample data acquisition: obtaining a large number of user credit samples, wherein the user credit samples include sample user credit characteristics and sample basic user credit score data;

[0124] Sample data labeling: Based on the needs of the scenario, the expert service is introduced to manually label the user credit samples with the corresponding sample labels. The score change factor data label is based on the score change factor data label of the user credit samples annotated by the service platform.

[0125] Model training process: input sample data into the initial score change factor model for at least one round of model training, predict the predicted score change factor data for the sample basic user credit score data through the initial score change factor model, determine the model loss value based on the predicted score change factor data and the score change factor data label using the model loss function, adjust the model parameters of the initial score change factor model based on the model loss value, and obtain the score change factor model by meeting the model training end conditions.

[0126] Optionally, the model training end condition of the model may include, for example, the value of the loss function is less than or equal to a preset loss function threshold, the number of iterations reaches a preset number threshold, etc. The specific model training end condition can be determined based on actual conditions and is not specifically limited here.

[0127] It should be noted that the machine learning models involved in one or more embodiments of this specification include but are not limited to convolutional neural network (CNN) models, deep neural network (DNN) models, recurrent neural network (RNN) models, embedding models, gradient boosting decision tree (GBDT) models, logistic regression (LR) models and other machine learning models. One or more fitting

[0128] Optionally, the model loss function may be a fit using one or more of the hinge loss calculation function, the cross entropy loss calculation function, the Euclidean distance loss calculation function, and the like in the related technology.

[0129] In one or more embodiments of this specification, the above method can be used to train a high-quality score change factor model, input the user's underlying attribute characteristics, credit indicator trend characteristics, and overdue indicator characteristics into the score change factor model, dynamically identify and determine the degree of correlation between the user's credit score change and the actual credit risk change, and generate confidence data as the basis for score adjustment. The accuracy and stability of the credit score are effectively improved, so that the score more truly reflects the user's credit risk changes, and the score instability caused by external environment or data fluctuations is reduced, thereby providing more reliable support for credit affairs decision-making and risk management, and reducing potential credit risks.

[0130] The following will be combined Figure 8 , the scoring processing device provided in this specification is introduced in detail. It should be noted that, Figure 8 The scoring processing device shown is used to execute the Figure 1 to Figure 7 For the convenience of explanation, only the parts related to this specification are shown. For the specific technical details not disclosed, please refer to this specification. Figure 1 to Figure 7 The embodiment shown.

[0131] See also Figure 8 , which shows a schematic diagram of the structure of the scoring processing device of this specification. The scoring processing device 1 can be implemented as all or part of the user electronic device through software, hardware or a combination of both. According to some embodiments, the scoring processing device 1 includes a basic scoring module 11, a factor determination module 12 and a scoring correction module 13, which are specifically used to:

[0132] A basic scoring module 11 is used to determine basic user credit scoring data for a target user based on a credit scoring model, and determine user credit characteristics for the target user;

[0133] A factor determination module 12 is used to determine score change factor data for the basic user credit score data using a score change factor model based on user credit characteristics, wherein the score change factor data is confidence level data that the score change factor model identifies the change in the basic user credit score to represent the change in the user's real credit risk;

[0134] The score correction module 13 is used to use a credit score correction model to perform score adjustment based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain a target user credit score for the target user, wherein the historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

[0135] Optionally, the factor determination module 12 is used to:

[0136] Determine the user's underlying attribute characteristics and the user's credit index trend characteristics from the user's credit characteristics, and determine the user's overdue status change characteristics based on the user's credit overdue information in the user's credit characteristics;

[0137] Based on the underlying attribute characteristics of the user, the trend characteristics of the user credit index, and the characteristics of the change in the overdue status of the user, a score change factor model is used to determine the basic user credit score change information and the user's real credit risk change information for the basic user credit score data;

[0138] Based on the basic user credit score change information and the user's real credit risk change information, a score change factor model is used to identify the confidence data of the basic user credit score change that represents the user's real credit risk change, and the confidence data is used as the score change factor data for the basic user credit score data.

[0139] Optionally, the factor determination module 12 is used to:

[0140] Determine a first credit overdue status at a first evaluation time and a second credit overdue status at a second evaluation time based on the user credit overdue information in the user credit feature, wherein the first evaluation time and the second evaluation time are separated by a preset time interval;

[0141] If the first credit overdue status matches the second credit overdue status, generating a user overdue status change feature of a stable status type;

[0142] If the first credit overdue status does not match the second credit overdue status, a user overdue status change feature of a status change type is generated.

[0143] Optionally, the score correction module 13 is used to:

[0144] Determine a basic user credit score time series from the basic user credit score data, and determine a score change factor time series from the score change factor data;

[0145] The basic user credit score time series and the score change factor time series are input into a credit score correction model, and the target basic user credit score is weightedly processed by the credit score correction model using a score weight parameter and historical user credit score data to obtain a target user credit score for the target user, wherein the target basic user credit score is the basic user credit score corresponding to the target score time unit in the basic user credit score data.

[0146] Optionally, the score correction module 13 is used to:

[0147] Inputting the basic user credit score time series and the score change factor time series into a credit score correction model, determining a score weight parameter between the basic user credit score item and the historical user credit score item through the credit score correction model, and obtaining the historical user credit score data corresponding to the score weight parameter and the historical user credit score item;

[0148] Obtaining a target user credit score using a target formula based on the target basic user credit score, the historical user credit score data, and the score weight parameter;

[0149] The target formula satisfies the following formula:

[0150] R=w*y+(1-w)*r

[0151] Among them, R represents the target user credit score, w represents the score weight parameter, r represents the historical user credit score data, and y is the target basic user credit score.

[0152] Optionally, the score correction module 13 is used to:

[0153] Inputting the basic user credit score time series and the score change factor time series into the credit score correction model to perform at least one round of weight learning training;

[0154] During the weight learning training process, the current scoring weight parameters and credit behavior labels are determined, and the predicted user credit score data is determined using the target formula based on the basic user credit score time series, the scoring weight parameters and the historical user credit score data. The weight learning loss is calculated based on the predicted user credit score data, the credit behavior label, the score change factor time series and the historical user credit score data, and the scoring weight parameters are adjusted using the weight learning loss until the weight learning training end condition is met.

[0155] Optionally, the score correction module 13 is used to:

[0156] Based on the predicted user credit score data, the credit behavior label, the score change factor time series and the historical user credit score data, a weighted learning loss calculation formula is used to obtain a weighted learning loss;

[0157] The weight learning loss calculation satisfies the following formula:

[0158]

[0159] Wherein, L represents the weight learning loss, T represents the upper limit of the time unit of the time series, t represents the tth time unit, N represents the number of time units of the time series, BCE() represents the cross entropy operation processing, and R t represents the predicted user credit score corresponding to the tth time unit in the predicted user credit score data, and the y t represents the credit behavior label value corresponding to the t-th time unit in the credit behavior label, and the λ t represents the fractional change factor value corresponding to the tth time unit in the fractional change factor time series, and the R t-1 Represents the predicted user credit score corresponding to the t-1th time unit in the predicted user credit score data.

[0160] Optionally, the device 1 is further used for:

[0161] Create an initial score change factor model;

[0162] Obtain user credit samples and label the score change factor data labels of the user credit samples, wherein the user credit samples include sample user credit features and sample basic user credit score data;

[0163] Based on the user credit sample, at least one round of model training is performed on the initial score change factor model, and predicted score change factor data for the sample base user credit score data is predicted by the initial score change factor model. Based on the predicted score change factor data and the score change factor data label, model parameters of the initial score change factor model are adjusted until the model training is completed to obtain a score change factor model.

[0164] It should be noted that the scoring processing device provided in the above embodiment only uses the division of the above functional modules as an example when executing the scoring processing method. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the scoring processing device provided in the above embodiment and the scoring processing method embodiment belong to the same concept, and the implementation process thereof is detailed in the method embodiment, which will not be repeated here.

[0165] The above serial numbers in this specification are for description only and do not represent the advantages or disadvantages of the embodiments.

[0166] In one or more embodiments of the present specification, the service platform first determines the basic user credit score data for the target user based on the credit score model, determines the score change factor data for the basic user credit score data based on the credit characteristics of the target user using the score change factor model, and then uses the credit score correction model to adjust the score based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain the target user credit score for the target user, thereby overcoming the disadvantage of the traditional credit score model in that the fluctuation of the underlying data leads to the fluctuation of the credit score of the individual user, and introducing the score change factor model to identify the score change factor corresponding to the actual risk change of the change in the user's credit score, and adjusting the basic user credit score according to the actual credit risk change of the user to obtain an accurate target user credit score, thereby ensuring the stability and rationality of the user's credit score, so that the user's credit score is more in line with the actual risk change, and can provide a more reliable basis for credit decision-making and risk management.

[0167] The present specification also provides a computer storage medium, which can store multiple instructions, which are suitable for being loaded and executed by a processor as described above. Figure 1 to Figure 7 The scoring method of the embodiment shown in the figure can be found in the specific implementation process. Figure 1 to Figure 7 The specific description of the illustrated embodiment will not be repeated here.

[0168] The present specification also provides a computer program product, which stores at least one instruction, and the at least one instruction is loaded and executed by the processor as described above. Figure 1 to Figure 7 The scoring method of the embodiment shown in the figure can be found in the specific implementation process. Figure 1 to Figure 7 The specific description of the illustrated embodiment will not be repeated here.

[0169] Please refer to Fig. 9 , is a block diagram of a structure of an electronic device provided in an embodiment of this specification. The electronic device in this specification may include one or more of the following components: a processor 1010, a memory 1020, an input device 1030, an output device 1040, and a bus 1050. The processor 1010, the memory 1020, the input device 1030, and the output device 1040 may be connected via a bus 1050.

[0170] The processor 1010 may include one or more processing cores. The processor 1010 uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions and processes data of the electronic device 100 by running or executing instructions, programs, code sets or instruction sets stored in the memory 1020, and calling data stored in the memory 1020. Optionally, the processor 1010 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 1010 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1010, but may be implemented separately through a communication chip.

[0171] The memory 1020 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 1020 includes a non-transitory computer-readable storage medium. The memory 1020 may be used to store instructions, programs, codes, code sets, or instruction sets.

[0172] The input device 1030 is used to receive input instructions or data, and the input device 1030 includes but is not limited to a keyboard, a mouse, a camera, a microphone or a touch device. The output device 1040 is used to output instructions or data, and the output device 1040 includes but is not limited to a display device and a speaker. In the embodiment of this specification, the input device 1030 can be a temperature sensor for obtaining the operating temperature of the electronic device. The output device 1040 can be a speaker for outputting audio signals.

[0173] In addition, those skilled in the art will appreciate that the structure of the electronic device shown in the above drawings does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown, or combine certain components, or arrange the components differently. For example, the electronic device also includes a radio frequency circuit, an input unit, a sensor, an audio circuit, a wireless fidelity (WIFI) module, a power supply, a Bluetooth module and other components, which will not be described in detail here.

[0174] In the embodiments of this specification, the execution subject of each step may be the electronic device described above. Optionally, the execution subject of each step is the operating system of the electronic device. The operating system may be an Android system, an IOS system, or other operating systems, which is not limited in the embodiments of this specification.

[0175] exist Fig. 9 In the electronic device, the processor 1010 can be used to call the program stored in the memory 1020 and execute it to implement the scoring processing method as described in the various method embodiments of this specification.

[0176] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0177] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in the embodiments of this specification are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions. For example, the basic user credit score data and target user credit characteristics involved in this specification are all obtained with full authorization.

[0178] The above disclosure is only the preferred embodiment of this specification, which certainly cannot be used to limit the scope of rights of this specification. Therefore, equivalent changes made according to the claims of this specification are still within the scope covered by this specification.

Claims

1. A scoring processing method, the method comprising: Determine basic user credit score data for the target user based on the credit score model, and determine user credit characteristics for the target user; Based on the user's credit characteristics, a score change factor model is used to determine score change factor data for the basic user's credit score data, wherein the score change factor data is confidence level data that the score change factor model identifies the change in the basic user's credit score to represent the change in the user's real credit risk; Based on the basic user credit score data, the score change factor data and the historical user credit score data, a credit score correction model is used to perform score adjustment to obtain a target user credit score for the target user. The historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

2. The method according to claim 1, wherein the step of determining the score change factor data for the basic user credit score data based on the user credit characteristics and using the score change factor model comprises: Determine the user's underlying attribute characteristics and the user's credit index trend characteristics from the user's credit characteristics, and determine the user's overdue status change characteristics based on the user's credit overdue information in the user's credit characteristics; Based on the underlying attribute characteristics of the user, the trend characteristics of the user credit index, and the characteristics of the change in the overdue status of the user, a score change factor model is used to determine the basic user credit score change information and the user's real credit risk change information for the basic user credit score data; Based on the basic user credit score change information and the user's real credit risk change information, a score change factor model is used to identify the confidence data of the basic user credit score change that represents the user's real credit risk change, and the confidence data is used as the score change factor data for the basic user credit score data.

3. The method according to claim 2, wherein determining the user overdue status change characteristics based on the user credit overdue information in the user credit characteristics comprises: Determine a first credit overdue status at a first evaluation time and a second credit overdue status at a second evaluation time based on the user credit overdue information in the user credit feature, wherein the first evaluation time and the second evaluation time are separated by a preset time interval; If the first credit overdue status matches the second credit overdue status, generating a user overdue status change feature of a stable status type; If the first credit overdue status does not match the second credit overdue status, a user overdue status change feature of a status change type is generated.

4. The method according to claim 1, wherein the step of adjusting the score by using a credit score correction model based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain a target user credit score for the target user comprises: Determine a basic user credit score time series from the basic user credit score data, and determine a score change factor time series from the score change factor data; The basic user credit score time series and the score change factor time series are input into a credit score correction model, and the target basic user credit score is weightedly processed by the credit score correction model using a score weight parameter and historical user credit score data to obtain a target user credit score for the target user, wherein the target basic user credit score is the basic user credit score corresponding to the target score time unit in the basic user credit score data.

5. The method according to claim 4, wherein the basic user credit score time series and the score change factor time series are input into a credit score correction model, and the target basic user credit score is weighted by the credit score correction model using a score weight parameter and historical user credit score data to obtain a target user credit score for the target user, comprising: Inputting the basic user credit score time series and the score change factor time series into a credit score correction model, determining a score weight parameter between the basic user credit score item and the historical user credit score item through the credit score correction model, and obtaining the historical user credit score data corresponding to the score weight parameter and the historical user credit score item; Obtaining a target user credit score using a target formula based on the target basic user credit score, the historical user credit score data, and the score weight parameter; The target formula satisfies the following formula: R=w*y+(1-w)*r Among them, R represents the target user credit score, w represents the score weight parameter, r represents the historical user credit score data, and y is the target basic user credit score.

6. The method according to claim 5, wherein the inputting the basic user credit score time series and the score change factor time series into a credit score correction model, and determining the score weight parameter between the basic user credit score item and the historical user credit score item through the credit score correction model, comprises: Inputting the basic user credit score time series and the score change factor time series into the credit score correction model to perform at least one round of weight learning training; During the weight learning training process, the current scoring weight parameters and credit behavior labels are determined, and the predicted user credit score data is determined using the target formula based on the basic user credit score time series, the scoring weight parameters and the historical user credit score data. The weight learning loss is calculated based on the predicted user credit score data, the credit behavior label, the score change factor time series and the historical user credit score data, and the scoring weight parameters are adjusted using the weight learning loss until the weight learning training end condition is met.

7. The method according to claim 6, wherein the step of calculating the weight learning loss based on the predicted user credit score data, the credit behavior label, the score change factor time series, and the historical user credit score data comprises: Based on the predicted user credit score data, the credit behavior label, the score change factor time series and the historical user credit score data, a weighted learning loss calculation formula is used to obtain a weighted learning loss; The weight learning loss calculation satisfies the following formula: Wherein, L represents the weight learning loss, T represents the upper limit of the time unit of the time series, t represents the tth time unit, N represents the number of time units of the time series, BCE() represents the cross entropy operation processing, and R t represents the predicted user credit score corresponding to the tth time unit in the predicted user credit score data, and the y t represents the credit behavior label value corresponding to the t-th time unit in the credit behavior label, and the λ t represents the fractional change factor value corresponding to the tth time unit in the fractional change factor time series, and the R t-1 Represents the predicted user credit score corresponding to the t-1th time unit in the predicted user credit score data.

8. The method according to claim 1, further comprising: Create an initial score change factor model; Obtaining user credit samples and labeling the score change factor data labels of the user credit samples, wherein the user credit samples include sample user credit features and sample basic user credit score data; Based on the user credit sample, at least one round of model training is performed on the initial score change factor model, and predicted score change factor data for the sample base user credit score data is predicted by the initial score change factor model. Based on the predicted score change factor data and the score change factor data label, model parameters of the initial score change factor model are adjusted until the model training is completed to obtain a score change factor model.

9. A scoring processing device, comprising: A basic scoring module, used to determine basic user credit score data for a target user based on a credit scoring model, and determine user credit characteristics for the target user; A factor determination module, for determining score change factor data for the basic user credit score data using a score change factor model based on user credit characteristics, wherein the score change factor data is confidence level data that the score change factor model identifies a change in the basic user credit score to represent a change in the user's real credit risk; A score correction module is used to use a credit score correction model to perform score adjustment based on the basic user credit score data, the score change factor data and the historical user credit score data to obtain a target user credit score for the target user, wherein the historical user credit score data is the historical user credit score output by the credit score correction model for the target user.

10. A computer storage medium storing a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the method steps according to any one of claims 1 to 8.

11. A computer program product, the computer program product storing at least one instruction, wherein the at least one instruction is loaded by a processor and executes the method steps according to any one of claims 1 to 8.

12. An electronic device comprising: A processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the method steps as claimed in any one of claims 1 to 8.

Images