Transaction risk assessment methods and devices, electronic equipment, and storage media

Abstract

This application discloses a method and apparatus for assessing transaction risk. The method includes: obtaining the customer group to which the user whose transaction risk is to be assessed belongs, wherein the user whose transaction risk is to be assessed is included in the customer group; obtaining a first risk score output by a risk deterministic model for each user in the customer group, wherein the risk deterministic model is constructed based on risk characteristics of a deterministic dimension; and obtaining a second risk score output by an uncertainty-addition model for each user, wherein the uncertainty-addition model is constructed based on risk characteristics of a non-deterministic dimension; obtaining customer group intervals with different risk levels based on the first and second risk scores of all users in the customer group; and determining the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among the multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs. This application can guarantee the accuracy of the transaction risk value.

Description

Technical Field

[0001] This application relates to the field of information security technology, and more specifically, to a transaction risk assessment method and apparatus, electronic device, and computer-readable storage medium. Background Technology

[0002] According to the guidelines of credit card organizations, when a disputed credit card transaction occurs, the cardholder can submit a chargeback request to the merchant's acquiring institution through the credit card issuer. If the chargeback request is accepted by the acquiring institution, the acquiring institution will cancel the relevant credit card transaction and refund the amount paid by the credit card to the cardholder through the issuing institution.

[0003] Every chargeback application accepted by an acquiring institution results in a certain degree of loss. Therefore, how to reduce the risk of credit card transaction chargebacks is an urgent problem to be solved. Summary of the Invention

[0004] To address the aforementioned technical problems, embodiments of this application provide a transaction risk assessment method and apparatus, an electronic device, and a computer-readable storage medium.

[0005] According to one aspect of the embodiments of this application, a method for assessing transaction risk is provided, comprising: obtaining a customer group to which a user whose transaction risk is to be assessed belongs, wherein the user whose transaction risk is to be assessed is included in the customer group; obtaining a first risk score output by a risk deterministic model for each user in the customer group, wherein the risk deterministic model is constructed based on risk characteristics of a deterministic dimension; and obtaining a second risk score output by an uncertainty-addition model for each user, wherein the uncertainty-addition model is constructed based on risk characteristics of a non-deterministic dimension; obtaining customer group intervals with different risk levels based on the first and second risk scores of all users in the customer group, wherein the risk level of the customer group interval is determined based on historical transaction information of multiple users in the customer group interval; and determining the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among the multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

[0006] According to one aspect of the embodiments of this application, a transaction risk assessment device is provided, comprising: a customer group acquisition module configured to acquire a customer group to which a user whose transaction risk is to be assessed belongs, wherein the user whose transaction risk is to be assessed is included in the customer group; a risk score acquisition module configured to acquire a first risk score output by a risk deterministic model for each user in the customer group, wherein the risk deterministic model is constructed based on risk characteristics of a deterministic dimension, and to acquire a second risk score output by an uncertainty-addition model for each user, wherein the uncertainty-addition model is constructed based on risk characteristics of a non-deterministic dimension; a customer group interval acquisition module configured to acquire customer group intervals with different risk levels based on the first and second risk scores of all users in the customer group, wherein the risk level of the customer group interval is determined based on historical transaction information of multiple users in the customer group interval; and a transaction risk acquisition module configured to determine the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among the multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

[0007] According to one aspect of the present application, an electronic device is provided, including a processor and a memory, wherein computer-readable instructions are stored in the memory, and when executed by the processor, the computer-readable instructions implement the transaction risk assessment method as described above.

[0008] According to one aspect of the embodiments of this application, a computer-readable storage medium is provided, on which computer-readable instructions are stored, which, when executed by a computer's processor, cause the computer to perform the transaction risk assessment method as described above.

[0009] According to one aspect of the embodiments of this application, a computer program product or computer program is provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the transaction risk assessment method provided in the various optional embodiments described above.

[0010] In the technical solution provided by the embodiments of this application, the transaction risk of a user is assessed based on the user's customer group. The assessment process incorporates both a risk deterministic model and an uncertainty-addition model. This assessment method obtains the user's transaction risk value by identifying differences in customer group risk. Acquiring institutions can quantify the user's transaction risk by assessing this value and handle risk mitigation for transactions initiated by the user's account, thereby reducing the risk of credit card chargebacks and minimizing losses for the acquiring institution.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0012] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0013] Figure 1 This is a flowchart illustrating a transaction risk assessment method as shown in an exemplary embodiment of this application;

[0014] Figure 2 This is an exemplary diagram illustrating changes in risk characteristics;

[0015] Figure 3 This is another exemplary diagram illustrating changes in risk characteristics;

[0016] Figure 4 yes Figure 1 A flowchart of step S150 in one embodiment is shown below;

[0017] Figure 5 yes Figure 1 A flowchart of step S170 in one embodiment is shown below;

[0018] Figure 6 This is an example of a customer group data diagram;

[0019] Figure 7 It is aimed at Figure 6 The diagram shows the distribution of transaction risk values ​​obtained by processing customer group data.

[0020] Figure 8 This is a block diagram illustrating a transaction risk assessment device in an exemplary embodiment of this application;

[0021] Figure 9 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0023] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0024] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0025] In this application, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0026] Currently, a mainstream risk scoring model system exists in the credit card transaction risk handling process: the FICO credit scoring system. This is a credit scoring system widely used in country X, which mainly assesses a user's creditworthiness from five dimensions: repayment history, number of credit accounts, length of time credit has been used, newly opened credit accounts, and types of credit currently in use. The credit score obtained from the FICO system typically ranges from 300 to 850 points; the higher the user's credit score, the lower their credit risk.

[0027] One key reason why the FICO credit scoring system is widely used in Country X is that it is a standardized and objective system built upon a vast collection of individual credit records and rigorous model refinement and stress testing. In other words, the FICO system relies on a large amount of historical data accumulated in the credit risk market, making it unsuitable for other countries or different business scenarios.

[0028] For example, in the business scenario of international payment bank card (hereinafter referred to as "foreign card", which refers to credit cards issued by international credit card organizations) transactions, currently, foreign card transactions only require the validity of the entered credit card number, expiration date, and credit card security code to proceed. The validity of most other information cannot be fully verified, and the validity of real-name information is extremely difficult to verify. Because foreign card acquiring institutions do not possess a large amount of accumulated information on user credit risk, transaction behavior, and real-name information comparison, they lack substantial and distinctive feature data, making it impossible to develop an effective transaction risk scoring model based on the FICO credit scoring system.

[0029] This application takes into account that the foreign card transaction business has a very wide range of customer groups, foreign card banks and initiating locations, and that the transaction risks of different customer groups in different regions are significantly different. In addition, the foreign card transaction business is currently in a period of transaction growth, and different customer groups may experience structural growth or changes in a certain region. Therefore, this application proposes a transaction risk assessment scheme that can accurately assess the transaction risk of users. For details, please refer to the content described in the following embodiments.

[0030] It should be noted that the transaction risk assessment scheme proposed in this application is not limited to the above-mentioned foreign card business scenario, but can also be applied to other types of credit card business scenarios. This embodiment does not limit this.

[0031] It should be noted that in the specific implementation of this application, user-related data is involved. When the embodiments of this application are applied to specific products or technologies, permission or consent from the user is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0032] Figure 1 This is a flowchart illustrating a transaction risk assessment method as shown in an exemplary embodiment of this application. Figure 1 As shown, this transaction risk assessment method includes at least steps S110 to S170, which are detailed below:

[0033] Step S110: Obtain the customer group to which the user whose transaction risk is to be assessed belongs. The user whose transaction risk is to be assessed is included in the customer group.

[0034] The user whose transaction risk is to be assessed mentioned in this embodiment should be understood as a user identifier, which is used to characterize the user whose transaction risk is to be assessed. The user identifier may be, for example, the credit card account name used by the user whose transaction risk is to be assessed, or other identifying information; this embodiment does not limit this.

[0035] A customer group refers to a set of users. The customer group to which a user whose transaction risk is to be assessed belongs can be determined based on the type of business the user uses to make transactions with a credit card. For example, users who make transactions of the same type of business can be included in the customer group, or users who have transaction anomalies of the same type of business can be included in the customer group. The customer group can also contain both types of users at the same time, and this embodiment does not limit this.

[0036] It should be noted that the types of transactions a user makes using a credit card can include gaming, e-commerce, live streaming, etc., and can be determined based on the actual application scenario; this document does not impose any restrictions. Users within the customer base who exhibit abnormal transactions within the same type of transaction can provide more significant risk characteristics for the subsequent transaction risk assessment process, thus improving the effectiveness of the transaction risk assessment.

[0037] It should also be noted that the transaction risks involved in this embodiment may include transaction chargeback risk, as well as other types of transaction risks. The specific risk type can be determined according to the actual business scenario.

[0038] Step S130: Obtain the first risk score output by the risk deterministic model for each user in the customer group. The risk deterministic model is constructed based on the risk characteristics of the deterministic dimension. Also obtain the second risk score output by the uncertainty addition model for each user. The uncertainty addition model is constructed based on the risk characteristics of the non-deterministic dimension.

[0039] Among them, the risk characteristics of the deterministic dimension refer to the characteristic information that can relatively objectively and stably assess the user's transaction risk. The risk characteristics of the deterministic dimension are not easily changed and do not involve frequent changes.

[0040] For example, in the context of foreign card transactions, certainty-based risk characteristics can include the country of the foreign card bank (different countries have different overall chargeback rates due to economic and other factors), the historical transaction records of the foreign card account (the risk profiles of old and new foreign card accounts differ), and the account's identity verification status (e.g., the identity verification of the ID card associated with the account, the identity verification of the passport associated with the account, etc.). These characteristics do not exhibit significant changes in risk levels with seasonality or the development of foreign card transaction business. Figure 2 As shown, the foreign card chargeback rates in developed country A and developing country B do not exhibit significant differences in risk with seasonality or the development of foreign card transaction business.

[0041] Certainty risk models can be constructed based on pre-collected risk characteristics of certain dimensions and data on risks occurring in actual business operations, such as building a logistic regression model. Therefore, certainty risk models built based on risk characteristics of certain dimensions typically have long model update cycles.

[0042] The risk characteristics of the uncertainty dimension refer to the features of user transaction risk that are difficult to assess relatively objectively and stably. The risk characteristics of the uncertainty dimension are prone to change and involve frequent fluctuations. Taking the foreign card transaction business scenario as an example, the risk characteristics of the uncertainty dimension can include changes in regional chargeback risk, regional structural changes (e.g., the region is in a business promotion period), changes in transaction methods, and changes in reimbursement channels. These characteristics may iterate rapidly with seasonal changes, product changes, etc.

[0043] Therefore, uncertainty-based models typically have short update cycles. For example, when the risk characteristics of the uncertainty dimension change, a new uncertainty model needs to be created promptly. Thus, uncertainty-based models must possess the characteristic of rapid iteration. Uncertainty-based models can be constructed using XGboost machine learning models, LightGBM machine learning models, and deep learning models.

[0044] For example Figure 3 As shown, if we assume that live streaming platform C has a strict review system in country A, with strict controls on recharge and other businesses, resulting in a relatively stable foreign card chargeback rate, but lax controls in country B, leading to frequent and fluctuating merchant policies in country B. If the uncertainty-addition model cannot iterate quickly, when live streaming merchants tighten policies in country B, the uncertainty-addition model cannot obtain a reasonable risk score, thus affecting the rationality of the transaction risk value assessed for users. When implementing risk response strategies based on users' transaction risk values, the transaction risk policy threshold cannot be relaxed in a timely and reasonable manner, blocking many users' legitimate transactions and resulting in a poor user experience. Conversely, when live streaming merchants relax policies in country B, the failure to promptly block risky transactions will lead to a surge in chargeback applications received by acquiring institutions, resulting in increased losses for acquiring institutions.

[0045] This embodiment demonstrates that the uncertainty-based model is effective for assessing transaction risks for new customer groups, while the certainty-based model is effective for assessing transaction risks for existing customer groups. The transaction risk assessment scheme proposed in this embodiment combines the risk scores output by both models for each user within the customer group to which the user's transaction risk is being assessed, thereby evaluating the user's transaction risk value. This approach exhibits strong adaptability to actual transaction business scenarios.

[0046] In this embodiment, the risk determinism model can obtain a user's first risk score based on the information provided by the user when initiating a transaction and some historical transaction information. The information provided by the user when initiating a transaction may include credit card number, expiration date, credit security code, etc. The uncertainty addition model can obtain a user's second risk score based on information such as user transaction characteristics, their customer group, and changes in the merchants they transact with.

[0047] Therefore, in this embodiment, each user in the customer group to which the user whose transaction risk is to be assessed belongs will receive a corresponding first risk score and a second risk score.

[0048] Step S150: Based on the first risk score and the second risk score of all users in the customer group, obtain customer group intervals with different risk levels. The risk level of the customer group interval is determined based on the historical transaction information of multiple users in the customer group interval.

[0049] This embodiment divides the customer group into multiple customer group intervals based on the first risk score and the second risk score of all users in the customer group to which the transaction risk to be assessed belongs. Different customer group intervals have different risk levels. A customer group interval can be understood as a two-dimensional interval formed by the first risk score and the second risk score. Each customer group interval contains multiple users, and each user can find a corresponding first risk score and second risk score in the customer group interval.

[0050] The risk level of a customer segment is determined based on the historical transaction information of multiple users within that segment. This information includes, for example, transaction risk details such as whether the user requested a chargeback. Because different users have different historical transaction records, the risk level varies across different customer segments.

[0051] It should be noted that the number of users in each customer segment may be the same or different, and the customer segment division can be based on actual needs. However, the number of users in each customer segment should not be too small, for example, it should not be less than a preset threshold, to ensure that the historical transaction information of users in each customer segment can accurately represent the transaction risk of that customer segment.

[0052] Step S170: Based on the risk level ranking among multiple customer group segments and the risk level of the customer group segment to which the user's transaction risk is to be assessed, determine the transaction risk value of the user whose transaction risk is to be assessed.

[0053] In this embodiment, the user's transaction risk value is determined based on the risk level of the customer group segment to which the user's transaction risk is to be evaluated, as well as the risk level ranking among multiple customer group segments. This is to evaluate the user's transaction risk by considering the overall transaction risk of the customer group, thereby improving the accuracy of the final transaction risk value.

[0054] For example, the risk level of the user's customer group segment to be assessed can be ranked in the risk level ranking, and the transaction risk value of the user to be assessed can be calculated based on the obtained ranking. Different rankings can use different methods to calculate the transaction risk value, so as to highlight the impact of the overall transaction risk of the customer group on the user's transaction risk. For details on the process of obtaining the transaction risk value, please refer to the following embodiments, which will not be repeated here.

[0055] Therefore, in the method proposed in this embodiment, the transaction risk of a user is assessed based on the customer group to which the user belongs. In the assessment process, a risk deterministic model and an uncertainty addition model are introduced, so that the transaction risk value obtained for the user whose transaction risk is to be assessed is obtained by differentiating customer group risk. This customer group risk differentiation identification method is also very applicable to the current credit card transaction risk situation, such as the transaction growth in the above-mentioned foreign card transaction business scenario, and can be adjusted accordingly according to the actual transaction risk environment. Therefore, the user transaction risk value obtained by the method proposed in this embodiment has extremely high accuracy.

[0056] Once a user's transaction risk value is assessed, appropriate risk mitigation measures can be taken based on that value. For example, if a user's transaction risk value exceeds a preset risk threshold, it indicates a high level of transaction risk. By limiting the user's maximum spending amount or the number of transactions, losses to the acquiring institution due to chargebacks can be significantly reduced. Specific risk mitigation methods can be determined based on the actual application scenario; this embodiment does not impose such limitations.

[0057] Figure 4 yes Figure 1 The flowchart of step S150 in the illustrated embodiment is shown in one embodiment. For example... Figure 4 As shown, step S150 describes the process of obtaining customer group intervals with different risk levels based on the first and second risk scores of all users in the customer group. This process includes steps S151 to S153, which are detailed below:

[0058] Step S151: A two-dimensional plane is formed by the first risk score and the second risk score of all users in the customer group.

[0059] This embodiment uses a two-dimensional scoring system to assess users' transaction risks, thereby ensuring the accuracy of user transaction risk assessment. This two-dimensional scoring system relies on risk scores output by a risk determinism model and an uncertainty-addition model as two scoring dimensions. Therefore, it is necessary to form a two-dimensional plane based on the first and second risk scores of all users in the customer group to which the user's transaction risk is to be assessed. Then, based on the formed two-dimensional plane, a two-dimensional scoring calculation is performed to obtain the transaction risk value of the user whose transaction risk is to be assessed.

[0060] The first risk score and second risk score of all users in the customer group can be sorted from largest to smallest to obtain a first risk score sequence and a second risk score sequence. Then, the first risk score sequence and the second risk score sequence are used as different dimensions of a two-dimensional plane to construct the two-dimensional plane. The dimensions of the two-dimensional plane include horizontal and vertical dimensions, which can also be understood as including the X-axis dimension and the Y-axis dimension.

[0061] The resulting two-dimensional plane is associated with users in the customer group through risk scores in two dimensions. It can be mapped to any user in the customer group through the first risk score and the second risk score in the two-dimensional plane. Therefore, it can be understood that all users in the customer group are distributed in the two-dimensional plane according to their respective first risk score and second risk score.

[0062] Step S153: Perform grid intersection processing in the two-dimensional plane to obtain grid regions with different risk levels in the two-dimensional plane. Grid regions with different risk levels correspond to customer group intervals with different risk levels.

[0063] This embodiment performs grid cross-processing in a two-dimensional plane to divide users within the associated customer groups into risk zones. Each grid region obtained through grid cross-processing contains multiple users; therefore, these grid regions correspond to different customer group intervals, and the users contained in the grid regions are also users within the corresponding customer group intervals. Since each user's transaction risk profile is different, each grid region should have a different risk level, meaning each customer group interval should have a different risk level.

[0064] For example, the process of performing mesh intersection processing in a two-dimensional plane to obtain mesh regions with different risk levels in the two-dimensional plane includes the following steps:

[0065] Based on the first and second risk score sequences corresponding to the two-dimensional plane, the two-dimensional plane is divided into multiple grid regions; based on the historical transaction information of multiple users contained in the grid regions, the risk level of the grid regions is calculated.

[0066] Regarding the division of the grid area, in some embodiments, the first risk score sequence and the second risk score sequence can be divided into multiple risk intervals of the same number, and the intersection of the intervals formed between the risk intervals is taken as the grid area in the two-dimensional plane. This embodiment does not limit the specific method of dividing the risk intervals corresponding to the first risk score sequence and the second risk score sequence. For example, each risk interval contains the first risk score or the second risk score corresponding to the average number of users, and each risk interval can also correspond to the average first risk score interval or the average second risk score interval. In practical applications, it can be determined according to actual needs.

[0067] In other embodiments, considering that the grid area obtained according to the above embodiments may contain a small number of users, and the risk level obtained based on the historical transaction information of a small number of users is not significant, affecting the overall transaction risk of the customer group, after dividing the first risk score sequence and the second risk score sequence into multiple risk intervals of the same number, the intersection of intervals with a user number less than or equal to a threshold is merged with the intersection of adjacent intervals based on the number of users within the intersection of the intervals, so that the number of users contained in the merged intersection of intervals is greater than the threshold. After the interval intersections are merged, the intersection of intervals contained in the two-dimensional plane is taken as the grid area in the two-dimensional plane.

[0068] Specifically, for intervals with user numbers less than or equal to a threshold, adjacent interval intersections can be searched in a two-dimensional plane. If an adjacent interval intersection with a risk level closest to the risk level of the intervals with user numbers less than or equal to the threshold is found, then the interval with user numbers less than or equal to the threshold is merged with the found adjacent interval intersection. If multiple adjacent interval intersections with the closest risk levels are found, then the adjacent interval intersection with the smallest area is selected from the found adjacent interval intersections and merged with the interval with user numbers less than or equal to the threshold. If multiple adjacent interval intersections with the smallest area are also found, then adjacent interval intersections in one direction are randomly selected for merging.

[0069] It should be noted that the intersection of adjacent intervals includes the intersections of intervals located above, below, to the left, and to the right of intervals in the two-dimensional plane where the number of users is less than a threshold. This method of merging interval intersections ensures that the risk level of the merged interval intersection does not change significantly compared to the original, minimizing the impact on the overall transaction risk of the customer group and improving the accuracy of the final transaction risk value to some extent.

[0070] To calculate the risk level of a grid area, historical transaction information of all users within the grid area is obtained. This historical transaction information is used to indicate whether there is transaction risk in the user's historical transactions. For example, based on the user's historical transaction information, it can be determined whether the user initiated a credit card chargeback application during the historical transaction process. The ratio between the number of users with transaction risk in historical transactions and the total number of users in the grid area is calculated, and this ratio can be used as the risk level of the grid area.

[0071] Since the risk level of a grid area is calculated based on the actual risk information of all users in that grid area during their historical transactions, it can accurately characterize the transaction risk level of the customer group segment corresponding to that grid area.

[0072] To facilitate understanding of the process of dividing users whose transaction risks are to be assessed into customer groups with different risk levels, the following is a detailed description of this process using a specific example:

[0073] Assuming the customer group to which the transaction risk is being assessed comprises 1000 users, i.e., 1000 user accounts, both the risk deterministic model and the uncertainty-additional model output a first risk score and a second risk score for each user. Since 100 of these users have initiated chargeback requests during their historical transactions, the transaction risk includes chargeback risk.

[0074] The first risk scores output by the risk deterministic model for all users are sorted from low to high to obtain a sequence of first risk scores. Risk intervals X1, X2, ..., X50 are obtained by dividing each interval into risk intervals of 20 users. By obtaining the maximum and minimum first risk scores for each risk interval, each risk interval can be represented as [0, aa1), [aa1, aa2), [aa2, aa3), ..., [aa49, 1). The same process is applied to the second risk scores output by the uncertainty model for all users, resulting in risk intervals Y1, Y2, ..., Y50, which are then represented using the second risk scores as [0, bb1), [bb1, bb2), [bb2, bb3), ..., [bb49, 1].

[0075] Based on the user information within the intersection of the risk intervals, the following information is obtained:

[0076] {[0, aa1), [0, bb1), (number of users who refused to pay, number of users who did not refuse to pay)}

[0077] {[aa1, aa2), [bb1, bb2), (number of users who refused to pay, number of users who did not refuse to pay)} ......

[0079] If, within a certain interval intersection, the number of users who refuse to pay plus the number of normal users is less than or equal to 5, then according to the above interval intersection merging strategy, the interval intersection is merged with the adjacent interval intersection. The risk level of the interval intersection is represented by the user chargeback rate within the interval intersection, which is the ratio between the number of users who refuse to pay and the total number of users.

[0080] Risk range ... [bb15, bb16) [bb16, bb17) [bb17, bb18) ... ... [aa18, aa19) (1，5) (2，5) (3，3) [aa19, aa20) (2，5) (2，2) (4，3) [aa20, aa21) (3，3) (2，5) (4，2) ...

[0081] Table 1

[0082] If we represent all the intersections of intervals as shown in Table 1 above, the intersections in the table represent the number of users who refused to pay and the number of users who did not refuse to pay within the interval intersections. It can be seen that the number of users in the interval intersections {[aa19, aa20), [bb16, bb17)} is less than 5, and the user chargeback rate is closest to that of the adjacent interval intersections {[aa19, aa20), [bb17, bb18)}. Therefore, we will merge the two interval intersections.

[0083] After merging all interval intersections where the number of users is less than a preset threshold of 5, the intersections of each interval can be used as grid areas in a two-dimensional plane, that is, as customer groups with different risk levels.

[0084] As can be seen from the above, the two-dimensional scoring system proposed in this embodiment employs a grid search strategy to divide the customer base into customer segments with different risk levels. This two-dimensional scoring system also includes a two-dimensional scoring calculation process to obtain the transaction risk value of the user whose transaction risk is to be assessed. For detailed two-dimensional scoring calculation procedures, please refer to [link to relevant documentation]. Figure 5 Corresponding implementation examples.

[0085] like Figure 5 As shown, Figure 1 The process of determining the transaction risk value of a user whose transaction risk is to be assessed based on the risk level ranking among multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs, as described in step S170 of the illustrated embodiment, includes at least steps S171 to S173, which are detailed below:

[0086] Step S171: Obtain the ranking of the risk level of the customer group interval to which the transaction risk to be assessed belongs in the risk level ranking.

[0087] By sorting the risk levels among multiple customer groups, the ranking of the risk level of the customer group to which the transaction risk is to be assessed is obtained.

[0088] Step S173: Calculate the transaction risk value of the user whose transaction risk is to be assessed based on the ranking.

[0089] If the user whose transaction risk is to be assessed is determined to be ranked at the lowest level, then the transaction risk value is calculated based on the first risk score and the second risk score corresponding to the user whose transaction risk is to be assessed, as well as the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

[0090] If the user whose transaction risk is to be assessed has a higher ranking than the lowest ranking, then the transaction risk value is calculated based on the first risk score and the second risk score of the user whose transaction risk is to be assessed, the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs, and the risk level of the customer group interval to which the user whose ranking is lower than the ranking of the user whose transaction risk is to be assessed.

[0091] For example, suppose that sorting multiple customer groups according to their risk level can result in a sorting pattern such as r1 < r2 < r3 < ..., where r1 represents the risk level of the lowest-ranked customer group. If the first risk score and the second risk score corresponding to the user whose transaction risk is to be assessed are represented as x1 and y1 respectively.

[0092] When the user whose trading risk is to be assessed corresponds to the lowest ranking, the trading risk value of the user whose trading risk is to be assessed is (x1) 2 +y1 2 The value is calculated as r1 to ensure that the user's trading risk value falls within the range of (0, r1); when the user's ranking is higher than the lowest ranking, for example, r... k (k≥2), the transaction risk value of the user whose transaction risk needs to be assessed is (x1) 2 +y1 2 )*(r k -r k-1 )+r k-1 .

[0093] As can be seen from the above, this application calculates the transaction risk value of the user whose transaction risk is to be assessed according to the ranking of the risk level of the customer group interval in the risk level ranking. The transaction risk value of the user in the high-risk area is assigned higher than that of the user in the low-risk area, which is consistent with the actual risk business scenario and makes the obtained transaction risk value highly accurate.

[0094] In another exemplary embodiment, customer group intervals with different risk levels are obtained multiple times based on the first and second risk scores of all users in the customer group to which the transaction risk to be assessed belongs. Each obtained customer group interval is generated according to a random risk interval division strategy. By executing this strategy, the customer group intervals obtained each time are made random.

[0095] Based on the customer group ranges with different risk levels obtained multiple times, the transaction risk value of the user to be assessed can be obtained multiple times accordingly. For details of the acquisition process, please refer to the aforementioned embodiment, which will not be repeated here.

[0096] The target transaction risk value is selected from multiple obtained transaction risk values ​​and used as the transaction risk value for the user to be evaluated. It should be noted that this embodiment does not restrict the specific method of selecting the target transaction risk value. A relatively reasonable transaction risk value with strong risk differentiation capability can be selected as the target transaction risk value according to actual application needs to further improve the accuracy of the final obtained transaction risk value. With a highly accurate transaction risk value, it is also possible to ensure that subsequent risk response measures for the user are effective.

[0097] To verify the significant effectiveness of the transaction risk assessment scheme proposed in this application, the inventors of this application adopted the following... Figure 6 A comparative experiment was conducted using the sample customer data, and the detailed process is as follows:

[0098] Based on the transaction risk assessment scheme proposed in this application, Figure 6 The customer group data shown is processed to obtain the interval intersection situation shown in Table 2 below. If a user has initiated a chargeback request in the historical transaction process, it is marked with "1", otherwise it is marked with "0". The intersection of the tables shown in Table 2 below represents the number of chargeback users and non-chargeback users in the customer group interval represented by the transaction interval.

[0099] Risk range [0.01,0.103] (0.103,0.185] (0.185,0.34] (0.34,0.77] [0.03,0.0775] (0,5) (0.0775,0.175] (1,2) (2,0) (0.175,0.31] (0,2) (0,3) (0.31,0.8] (3,2)

[0100] Table 2

[0101] According to the transaction risk assessment scheme proposed in this application, the information obtained from the intersection of the intervals shown in Table 2 above can be used to obtain the following: Figure 7 The distribution of transaction risk values ​​among all users in the customer group shown. Figure 7 The first risk interval shown refers to the first risk score sequence formed by the first risk scores of all users, and the second risk interval refers to the second risk score sequence formed by the second risk scores of all users. Figure 7 The risk level shown refers to the user chargeback rate within the intersection of the intervals.

[0102] according to Figure 7A sequence of customer groups ordered from lowest to highest risk based on user chargeback rate can be obtained: {[0.03,0.0775], [0.01,0.103]} = {(0.0175,0.31], (0.103,0.185]} = {(0.175,0.31], (0.185,0.34]} < {(0.0775,0.175], (0.103,0.185]} < {(0.31,0.8], (0.34,0.77]} < {(0.0775,0.175], (0.185,0.34]}, with risk levels ranked as 0 < 0.333 < 0.6 < 1.

[0103] The KS (Kolmogorov-Smirnov) evaluation method, which measures the model's ability to distinguish between positive and negative samples (a higher value indicates a better model performance), yields the following KS values ​​for the proposed transaction risk assessment scheme, the scheme using only a risk deterministic model to assess user transaction risk, and the scheme using only an uncertainty-based model to assess user transaction risk: Table 3 shows the corresponding KS values ​​for each of the following schemes:

[0104] This application Risk certainty model New models for uncertainty KS value 78.60％ 42.90％ 50.00％

[0105] Table 3

[0106] It should be noted that other methods can also be used to evaluate the effectiveness of the above three transaction risk assessment schemes, such as the inflow GINI assessment method, and this section does not impose any restrictions on this.

[0107] As can be seen from the above, the transaction risk assessment scheme proposed in this application has a very high KS value, which indicates that the transaction risk assessment scheme proposed in this application has a very significant effect on assessing the user's transaction risk value.

[0108] It should also be noted that the transaction risk assessment scheme proposed in this application can be deployed on any terminal or server to assess the user's transaction risk value. The specific deployment can be determined according to the actual application situation.

[0109] The term "terminal" as used herein can refer to any electronic device capable of running video playback clients, such as smartphones, tablets, laptops, and computers. The term "server" as used herein can refer to a standalone physical server, a server cluster or distributed system composed of multiple physical servers, where multiple servers can form a blockchain, and the server is a node on the blockchain. A server can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms; this is not a limitation.

[0110] Figure 8 This is a block diagram illustrating a transaction risk assessment device as shown in an exemplary embodiment of this application. Figure 8 As shown, the device includes:

[0111] The customer group acquisition module 210 is configured to acquire the customer group to which the user whose transaction risk is to be assessed belongs, and the user whose transaction risk is to be assessed is included in the customer group; the risk score acquisition module 230 is configured to acquire the first risk score output by the risk deterministic model for each user in the customer group, the risk deterministic model is constructed based on the risk characteristics of the deterministic dimension, and to acquire the second risk score output by the uncertainty addition model for each user, the uncertainty addition model is constructed based on the risk characteristics of the uncertainty dimension; the customer group interval acquisition module 250 is configured to acquire customer group intervals with different risk levels based on the first risk score and the second risk score of all users in the customer group, the risk level of the customer group interval is determined based on the historical transaction information of multiple users in the customer group interval; the transaction risk acquisition module 270 is configured to determine the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

[0112] In another exemplary embodiment, the customer segment acquisition module 250 includes:

[0113] The two-dimensional plane forming unit is configured to form a two-dimensional plane from the first risk score and the second risk score of all users in the customer group; the grid area acquisition unit is configured to perform grid intersection processing in the two-dimensional plane to obtain grid areas with different risk levels in the two-dimensional plane, and the grid areas with different risk levels correspond to customer group intervals with different risk levels.

[0114] In another exemplary embodiment, the two-dimensional planar forming unit includes:

[0115] The risk score sorting subunit is configured to sort the first risk score and the second risk score of all users in the customer group from smallest to largest, to obtain the first risk score sequence and the second risk score sequence; the two-dimensional plane construction subunit is configured to construct the two-dimensional plane by using the first risk score sequence and the second risk score sequence as different dimensions of the two-dimensional plane.

[0116] In another exemplary embodiment, the grid region acquisition unit includes:

[0117] The grid area division sub-unit is configured to divide the two-dimensional plane into multiple grid areas based on the first risk score sequence and the second risk score sequence corresponding to the two-dimensional plane; the risk level calculation sub-unit is configured to calculate the risk level of the grid area based on the historical transaction information of multiple users contained in the grid area.

[0118] In another exemplary embodiment, the grid region partitioning subunit includes:

[0119] The risk interval division subunit is configured to divide the first risk score sequence and the second risk score sequence into multiple risk intervals of the same number; the interval intersection merging subunit is configured to merge the interval intersections with the number of users less than or equal to the number threshold with the adjacent interval intersections based on the number of users in the interval intersections between risk intervals, so that the number of users contained in the merged interval intersections is greater than the number threshold; the grid area acquisition subunit is configured to use the interval intersections formed in the two-dimensional plane as the grid area in the two-dimensional plane.

[0120] In another exemplary embodiment, the interval intersection union sub-unit includes:

[0121] The vector interval intersection search subunit is configured to search for adjacent interval intersections for intervals where the number of users is less than a threshold. The first merging subunit is configured to merge the interval intersection where the risk level is closest to that of the interval intersection where the number of users is less than the threshold with the searched adjacent interval intersection if the risk level is found. The second merging subunit is configured to select the adjacent interval intersection with the smallest interval area from the searched adjacent interval intersections and merge it with the interval intersection where the number of users is less than the threshold if there are multiple adjacent interval intersections.

[0122] In another exemplary embodiment, the risk level calculation subunit includes:

[0123] The historical transaction information acquisition subunit is configured to acquire historical transaction information of all users within the grid area. The historical transaction information is used to indicate whether there is transaction risk in the user's historical transactions. The risk level acquisition subunit is configured to calculate the ratio between the number of users with transaction risk in historical transactions and the total number of users within the grid area. The ratio is used as the risk level of the grid area.

[0124] In another exemplary embodiment, the transaction risk acquisition module 270 includes:

[0125] The ranking acquisition unit is configured to acquire the ranking of the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs in the risk level ranking; the transaction risk value calculation unit is configured to calculate the transaction risk value of the user whose transaction risk is to be assessed based on the ranking.

[0126] In another exemplary embodiment, the transaction risk value calculation unit includes:

[0127] The first calculation subunit is configured to calculate the transaction risk value based on the first and second risk scores of the user whose transaction risk is to be assessed, and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs, if the user whose transaction risk is to be assessed is determined to be at the lowest ranking. The second calculation subunit is configured to calculate the transaction risk value based on the first and second risk scores of the user whose transaction risk is to be assessed, the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs, and the risk level of the customer group interval to which the user whose ranking is lower than the ranking of the user whose transaction risk is to be assessed.

[0128] In another exemplary embodiment, the device further includes:

[0129] The customer segment generation module is configured to generate multiple customer segment intervals with different risk levels based on the first and second risk scores of all users in the customer segment. Each customer segment interval is generated according to a random risk segmentation strategy. The transaction risk value calculation module is configured to obtain the transaction risk value of the user whose transaction risk is to be evaluated multiple times based on the multiple customer segment intervals with different risk levels. The transaction risk value selection module is configured to select the target transaction risk value as the transaction risk value of the user whose transaction risk is to be evaluated from the multiple obtained transaction risk values.

[0130] It should be noted that the apparatus provided in the above embodiments and the method provided in the above embodiments belong to the same concept, and the specific way in which each module and unit performs operations has been described in detail in the method embodiments, and will not be repeated here.

[0131] Embodiments of this application also provide an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the transaction risk assessment method as described above.

[0132] Figure 9 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0133] It should be noted that, Figure 9 The computer system 1600 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0134] like Figure 9 As shown, the computer system 1600 includes a Central Processing Unit (CPU) 1601, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 1602 or programs loaded from storage portion 1608 into Random Access Memory (RAM) 1603, such as performing the methods described in the above embodiments. Various programs and data required for system operation are also stored in RAM 1603. The CPU 1601, ROM 1602, and RAM 1603 are interconnected via bus 1604. An Input / Output (I / O) interface 1605 is also connected to bus 1604.

[0135] The following components are connected to I / O interface 1605: an input section 1606 including a keyboard, mouse, etc.; an output section 1607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1608 including a hard disk, etc.; and a communication section 1609 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 1609 performs communication processing via a network such as the Internet. A drive 1610 is also connected to I / O interface 1605 as needed. Removable media 1611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1610 as needed so that computer programs read from them can be installed into storage section 1608 as needed.

[0136] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1609, and / or installed from removable medium 1611. When the computer program is executed by central processing unit (CPU) 1601, it performs various functions defined in the system of this application.

[0137] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0138] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0139] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0140] Another aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the transaction risk assessment method as described above. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0141] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the transaction risk assessment method provided in the various embodiments described above.

[0142] The above description is merely a preferred exemplary embodiment of this application and is not intended to limit the implementation of this application. Those skilled in the art can easily make corresponding modifications or alterations based on the main concept and spirit of this application. Therefore, the scope of protection of this application should be determined by the scope of protection claimed in the claims.

Claims

1. A method for assessing transaction risk, characterized in that, include: Obtain the customer group to which the user whose transaction risk is to be assessed belongs, wherein the user whose transaction risk is to be assessed is included in the customer group; The system obtains a first risk score from a risk deterministic model for each user in the customer group. The risk deterministic model is constructed based on risk characteristics of a deterministic dimension, which refers to characteristic information that can relatively objectively and stably assess transaction risk and is not easily affected by changes in transaction risk. The system also obtains a second risk score from an uncertainty-addition model for each user. The uncertainty-addition model is constructed based on risk characteristics of a non-deterministic dimension, which refers to characteristic information that is difficult to relatively objectively and stably assess transaction risk and is easily affected by changes in transaction risk. Based on the first risk score and the second risk score of all users in the customer group, customer group intervals with different risk levels are obtained, and the risk level of the customer group interval is determined based on the historical transaction information of multiple users in the customer group interval. Based on the risk level ranking among the multiple customer group segments and the risk level of the customer group segment to which the user whose transaction risk is to be assessed belongs, the transaction risk value of the user whose transaction risk is to be assessed is determined.

2. The method according to claim 1, characterized in that, The step of obtaining customer group intervals with different risk levels based on the first and second risk scores of all users in the customer group includes: A two-dimensional plane is formed by the first risk score and the second risk score of all users in the customer group; Grid intersection processing is performed in the two-dimensional plane to obtain grid regions with different risk levels in the two-dimensional plane, and the grid regions with different risk levels correspond to the customer group intervals with different risk levels.

3. The method according to claim 2, characterized in that, The two-dimensional plane formed by the first risk score and the second risk score of all users in the customer group includes: The first risk score and the second risk score of all users in the customer group are sorted from smallest to largest to obtain the first risk score sequence and the second risk score sequence; The first risk score sequence and the second risk score sequence are used as different dimensions of the two-dimensional plane to construct the two-dimensional plane.

4. The method according to claim 2, characterized in that, The step of performing mesh intersection processing in the two-dimensional plane to obtain mesh regions with different risk levels in the two-dimensional plane includes: Based on the first risk score sequence and the second risk score sequence corresponding to the two-dimensional plane, the two-dimensional plane is divided into multiple grid regions; The risk level of the grid area is calculated based on the historical transaction information of multiple users contained in the grid area.

5. The method according to claim 4, characterized in that, The step of dividing the two-dimensional plane into multiple grid regions based on the first risk score sequence and the second risk score sequence corresponding to the two-dimensional plane includes: The first risk score sequence and the second risk score sequence are divided into multiple risk intervals of the same number; Based on the number of users within the intersection of the risk intervals, the intersections of intervals with a number of users less than or equal to a threshold are merged with the intersections of adjacent intervals, so that the number of users contained in the merged intersection is greater than the threshold. The intersection of intervals formed in the two-dimensional plane is taken as the grid region in the two-dimensional plane.

6. The method according to claim 5, characterized in that, The step of merging the intersection of intervals with user numbers less than a threshold with the intersection of adjacent intervals includes: For intervals where the number of users is less than a threshold, search for the intersection of adjacent intervals. If the search finds an adjacent interval whose risk level is closest to the intersection of the intervals where the number of users is less than the threshold, then the intersection of the intervals where the number of users is less than the threshold is merged with the searched adjacent interval intersection.

7. The method according to claim 6, characterized in that, The method further includes: If there are multiple intersections of adjacent intervals found, the intersection with the smallest interval area is selected from the intersections of adjacent intervals found and merged with the intersection of intervals where the number of users is less than the number threshold.

8. The method according to claim 4, characterized in that, The step of calculating the risk level of the grid area based on the historical transaction information of multiple users contained in the grid area includes: Obtain historical transaction information for all users within the grid area; the historical transaction information is used to indicate whether there is any transaction risk in the user's historical transactions. The ratio between the number of users with transaction risks in historical transactions and the total number of users in the grid area is calculated, and the ratio is used as the risk level of the grid area.

9. The method according to claim 1, characterized in that, The step of determining the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among the multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs includes: Obtain the ranking of the risk level of the customer group segment to which the transaction risk to be assessed belongs in the risk level ranking; The transaction risk value of the user whose transaction risk is to be assessed is calculated based on the ranking.

10. The method according to claim 9, characterized in that, The calculation of the transaction risk value of the user to be assessed based on the ranking includes: If the user whose transaction risk is to be assessed is determined to be ranked at the lowest, then the transaction risk value is calculated based on the first risk score and the second risk score corresponding to the user whose transaction risk is to be assessed, as well as the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

11. The method according to claim 9, characterized in that, The calculation of the transaction risk value of the user to be assessed based on the ranking includes: If it is determined that the ranking of the user whose transaction risk is to be assessed is higher than the lowest ranking, then the transaction risk value is calculated based on the first risk score and the second risk score of the user whose transaction risk is to be assessed, the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs, and the risk level of the customer group interval to which the ranking is lower than the ranking of the user whose transaction risk is to be assessed.

12. The method according to claim 1, characterized in that, The method further includes: Based on the first risk score and the second risk score of all users in the customer group, customer group intervals with different risk levels are obtained multiple times. Each obtained customer group interval is generated according to a random risk interval division strategy. Based on multiple customer groups with different risk levels, the transaction risk value of the user whose transaction risk is to be assessed is obtained multiple times. The target transaction risk value is selected from the multiple obtained transaction risk values ​​as the transaction risk value of the user to be evaluated.

13. A transaction risk assessment device, characterized in that, include: The customer acquisition module is configured to acquire the customer group to which the user whose transaction risk is to be assessed belongs, wherein the user whose transaction risk is to be assessed is included in the customer group; The risk score acquisition module is configured to acquire a first risk score output by a risk deterministic model for each user in the customer group. The risk deterministic model is constructed based on risk characteristics of a deterministic dimension, which refers to characteristic information that can relatively objectively and stably assess transaction risk and is not easily affected by changes in transaction risk. The module is also configured to acquire a second risk score output by an uncertainty-added model for each user. The uncertainty-added model is constructed based on risk characteristics of a non-deterministic dimension, which refers to characteristic information that is difficult to relatively objectively and stably assess transaction risk and is easily affected by changes in transaction risk. The customer segment acquisition module is configured to acquire customer segment intervals with different risk levels based on the first risk score and the second risk score of all users in the customer segment, wherein the risk level of the customer segment interval is determined based on the historical transaction information of multiple users in the customer segment. The transaction risk acquisition module is configured to determine the transaction risk value of the user whose transaction risk is to be assessed based on the risk level ranking among the multiple customer group intervals and the risk level of the customer group interval to which the user whose transaction risk is to be assessed belongs.

14. An electronic device, characterized in that, include: Memory, which stores computer-readable instructions; A processor reads computer-readable instructions stored in memory to perform the method described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, It stores computer-readable instructions that, when executed by a computer's processor, cause the computer to perform the method described in any one of claims 1-12.

16. A computer program product, characterized in that, The computer program product includes a computer program stored in a computer-readable storage medium, wherein a processor of a computer device reads from the computer-readable storage medium and executes the computer program, causing the computer device to perform the method of any one of claims 1-12.

Images