A bank risk control model data sharing method and system

Abstract

The application provides a bank risk control model data sharing method and system, the method comprises the following steps: when a bank risk control model needs to register a public data table, configuring the public data table according to the bank risk control model, and configuring parameters of the bank risk control model; the public data table comprises data shared among the bank risk control models; when the bank risk control model is executed, reading and writing data in the public data table according to the configuration of the public data table, and calling the data in the public data table according to a calling rule. The method guarantees the consistency of the data, reduces the design work of data reuse logic among the bank risk control models, and improves the data sharing capability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of data sharing, and particularly relates to a bank risk control model inter-data sharing method and system. BACKGROUND

[0002] Now the bank risk control system is mature, and the bank risk control model construction is more and more, and is more and more complex, and the data reuse between multiple bank risk control models is inevitable. When multiple bank risk control models concurrently access or modify the same data source referenced by the multiple bank risk control models, the problem of inconsistent data and data accuracy may exist.

[0003] For example, referring to Figure 1 , suppose that there are a data table tbA, a model mdX, a model mdY and a model mdZ, the data table tbA contains a field of 10,000 yuan of foreign currency occurrence, and the three models are executed in sequence. The model mdX and the model mdY respectively convert the foreign currency occurrence into RMB, and then store the converted RMB into the field rmbBalance, and the model mdZ queries and accesses the value of the field rmbBalance. Assuming that the model mdX uses the intermediate price of the exchange rate of the day as 1.5 when converting, then the value of the field rmbBalance is 15,000 yuan; the model mdY uses the selling price 1.6 when converting RMB, so the value of the field rmbBalance at this time is 16,000 yuan; when the model mdZ accesses the field rmbBalance, if the execution time of the model is not well controlled, the data accessed by the model mdZ cannot be guaranteed to be converted according to which value, so that the execution result of the model cannot be guaranteed to be consistent and predictable under the premise of the same data source and algorithm.

[0004] In view of the above situation, the existing method solves the problem by using data sharing, and the existing bank risk control model inter-data sharing mainly uses the following two ways:

[0005] 1. Building a unified source data. The data is written and changed through a fixed system or a function module, and the bank risk control model has only read-only permission, reads data from the source data table, and establishes its own private table. The bank risk control model internally sets the data read-write permission and execution result identifier of the private table. The bank risk control model can freely use the private table, avoiding the problem of data confusion caused by the data being called by other bank risk control models, and ensuring the accuracy of the data. However, the data of the private table is the data processed by the bank risk control models, and the reuse utilization rate is low, the repeated calculation is much, and the storage pressure is large.

[0006] 2. Public data layer management. Adopt multi-layer public data source management to manage the data processed by the bank risk control model. For example, when the monthly cumulative amount of the account is needed to be calculated from the daily transaction details, the transaction details are the first layer, the daily account occurrence amount statistics are the second layer, and the monthly cumulative amount is the third layer. This method can make full use of public data, but lacks unified control mechanism, and the following risks exist: a. The read and write of each table in the multi-layer public data source lack permission control, and when multiple write-in occurs, data confusion will occur. b. The data access conditions, time points and update identifiers are mostly written by the bank risk control model, and the workload of the bank risk control model design is large, repetitive and prone to omissions.

[0007] In summary, controlling the consistency of public data between bank risk control models is the premise to ensure the correctness of the model execution result, and providing a perfect data sharing method is of great significance to the design of the bank risk control model. SUMMARY

[0008] In view of the defects in the prior art, the present application provides a bank risk control model data sharing method and system, which ensures the consistency of the data, reduces the design work of the data reuse logic between the bank risk control models, and improves the data sharing capability.

[0009] In the first aspect, a bank risk control model data sharing method comprises:

[0010] When a bank risk control model needs to operate a public data table, the bank risk control model is registered, the public data table is configured according to the bank risk control model, and the parameters of the bank risk control model are configured; the public data table comprises data shared between the bank risk control models;

[0011] When the bank risk control model is executed, the data in the public data table is read and written according to the configuration of the public data table, and the data in the public data table is called according to the calling rule.

[0012] Further, the bank risk control model is divided into a data owner, an attribute maintainer and a read-only access according to the relationship with the public data table;

[0013] The data owner is the model to which the data in the public data table belongs;

[0014] The attribute maintainer is the model that maintains the attribute field in the public data table;

[0015] The read-only access is the model that can only read the data in the public data table.

[0016] Further, the data owner configures the public data table, which specifically comprises:

[0017] Configuring the basic information in the public data table;

[0018] configure the attribute field to be maintained by the attribute maintainer;

[0019] give the data modification permission of the public data table, the data modification permission including the permission of modifying the data in the public data table;

[0020] configure the number of times of calling the data owner to be 1.

[0021] Further, the configuration of the public data table by the attribute maintainer specifically includes:

[0022] configure the public data table to be maintained;

[0023] configure the attribute field to be maintained;

[0024] check whether the maintenance frequency of the attribute field matches the upstream data.

[0025] Further, the configuration of the public data table by the read-only visitor specifically includes:

[0026] configure the public data table to be read;

[0027] check whether the data among the public data tables is synchronized.

[0028] Further, the bank risk control model is divided into real-time model, post-event model and quasi-real-time model according to time effectiveness;

[0029] The data updating mode of the real-time model is: updating in real time by receiving data;

[0030] The data updating mode of the post-event model is: updating by batch importing data in a long term;

[0031] The data updating mode of the quasi-real-time model is: updating by data synchronization in a short term.

[0032] Further, the calling rule of the post-event model includes:

[0033] obtain the data to be called;

[0034] record the preparation completion information of the data;

[0035] When it is judged according to the preparation completion information that the data is not ready, continue to wait for the preparation of the data;

[0036] When it is judged according to the preparation completion information that the data is ready, read all the data.

[0037] In a second aspect, a bank risk control model data sharing system includes:

[0038] The configuration unit is configured to configure the public data table according to the bank risk control model when the bank risk control model is registered, and configure parameters of the bank risk control model; the public data table includes data shared between the bank risk control models;

[0039] The execution unit is configured to read and write data in the public data table according to the configuration of the public data table, and call data in the public data table according to the calling rule when the bank risk control model is executed.

[0040] Further, the bank risk control model is divided into a data owner, an attribute maintainer and a read-only access according to the relationship with the public data table;

[0041] The data owner is a model to which data in the public data table belongs;

[0042] The attribute maintainer is a model to maintain attribute fields in the public data table;

[0043] The read-only access is a model that can only read data in the public data table;

[0044] The data owner is configured to specifically include:

[0045] Configure basic information in the public data table;

[0046] Configure attribute fields for the attribute maintainer to maintain;

[0047] Give data modification permission of the public data table, the data modification permission including permission to modify data in the public data table;

[0048] Configure the number of times the data owner is called to be 1;

[0049] The attribute maintainer is configured to specifically include:

[0050] Configure the public data table to be maintained;

[0051] Configure the attribute field to be maintained;

[0052] Check whether the maintenance frequency of the attribute field matches the upstream data;

[0053] The read-only access is configured to specifically include:

[0054] Configure the public data table to be read;

[0055] Check whether data between the public data tables is synchronized.

[0056] Further, the bank risk control model is divided into a real-time model, a post-event model and a quasi-real-time model according to the time effectiveness;

[0057] The data updating manner of the real-time model is to update in real time;

[0058] The data updating manner of the post-model is to update by batch importing data in a long term;

[0059] The data updating manner of the quasi-real-time model is to update by data synchronization in a short term;

[0060] The calling rule of the post-model comprises:

[0061] Obtaining data to be called;

[0062] Recording preparation completion information of the data;

[0063] When it is judged according to the preparation completion information that the data is not ready, continuing to wait for the preparation of the data;

[0064] When it is judged according to the preparation completion information that the data is ready, reading all the data.

[0065] It can be known from the above technical solution that the bank risk control model data sharing method and system provided by the application uniformly manages the bank risk control model and the public data table, guarantees the consistency of the data, reduces the design work of the data reuse logic between the bank risk control models, and improves the data sharing capability. The method completes the calling and reading and writing of the data according to the calling rule and the configuration of the public data table, and guarantees the uniformity of the timeliness of the data reuse. BRIEF DESCRIPTION OF DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the specific embodiments or the prior art, the drawings needed to be used in the specific embodiments or the prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual proportion.

[0067] Figure 1 The schematic diagram of the multi-model data access scene provided for the background art.

[0068] Figure 2 The flowchart of the bank risk control model data sharing method provided for the embodiment.

[0069] Figure 3 The schematic diagram of the bank risk control model configuration provided for the embodiment.

[0070] Figure 4 The schematic diagram of the post-model execution method provided for the embodiment.

[0071] Figure 5 The schematic diagram of the configuration and execution method illustrated for the embodiment.

[0072] Figure 6 The data sharing method provided for the embodiments is applied to Figure 1 a schematic diagram in a scenario. DETAILED DESCRIPTION

[0073] The embodiments of the technical solutions of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and therefore only serve as examples, and cannot limit the protection scope of the present application. It should be noted that, unless otherwise specified, the technical terms or scientific terms used in the present application should be understood as the usual meaning understood by the skilled person in the field of the present application.

[0074] It should be understood that, as used in this specification and the appended claims, the terms "comprises" and "comprising" indicate the presence of the described features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0075] It should also be understood that the terms used herein in the present application specification are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the present application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0076] As used in this specification and the appended claims, the term "if" can be interpreted as meaning "when" or "once" or "in response to a determination" or "in response to detecting" depending on the context. Similarly, the phrase "if it is determined" or "if [a described condition or event] is detected" can be interpreted as meaning "once it is determined" or "in response to the determination" or "once [a described condition or event] is detected" or "in response to detecting [a described condition or event]" depending on the context.

[0077] Embodiment one:

[0078] A bank risk control model data sharing method, see Figure 2 , comprising:

[0079] S1: When a bank risk control model needs to register to operate a public data table, configure the public data table according to the bank risk control model, and configure the parameters of the bank risk control model; the public data table includes data shared among bank risk control models;

[0080] S2: When the bank risk control model is executed, read and write data in the public data table according to the configuration of the public data table, and call data in the public data table according to the calling rule.

[0081] In the embodiment, the configuration of the parameters in the bank risk control model can include configuring the parameters of the entity, index, and the like of the bank risk control model, such as configuring the calling method of the bank risk control model (including being called by other models, being called by a scheduling system, and the like), the model time limit (including real-time, quasi-real-time, post-event, and the like), and the upper calling model, and the like. In the method, if the bank risk control model needs to operate the data in the public data table, the public data table needs to be configured, such as configuring the read-write permission of the public data table. The public data table can be multiple, and each public data table includes multiple data that need to be shared among the bank risk control models. The method can also query the data reuse situation through the public data table after the configuration of the public data table is completed. The method first registers the public data table in the bank risk control model registration stage, which simplifies the design work of data sharing in the later stage.

[0082] In the embodiment, the method completes the writing, deleting, maintaining, and accessing of the data in the public data table according to the configuration of the public data table in the model execution stage, and calls the data in the public data table according to the calling rule. The calling rule can be set according to different bank risk control models, and the calling rule can be set in advance or during the use of the method, which ensures the uniformity of the data calling time limit.

[0083] The method uniformly manages the bank risk control model and the public data table, ensures the consistency of the data, reduces the design work of the data reuse logic among the bank risk control models, and improves the data sharing capability. The method completes the calling and reading and writing of the data according to the calling rule and the configuration of the public data table, and ensures the uniformity of the data reuse time limit.

[0084] Further, in some embodiments, the bank risk control model is divided into a data owner, an attribute maintainer, and a read-only access according to the relationship with the public data table.

[0085] The data owner is the model to which the data in the public data table belongs.

[0086] The attribute maintainer is the model that maintains the attribute field in the public data table.

[0087] The read-only access is the model that can only read the data in the public data table.

[0088] In the embodiment, referring to Figure 3 The method divides the bank risk control model into three types according to the relationship with the public data table: a data owner, an attribute maintainer, and a read-only access. The data owner is the model to which the data in the public data table belongs, and has the unique data writing permission and data modification permission of the public data table. In this way, the data in the public data table is avoided from being modified by other models (non-data owners), causing the data in the public data table to be modified multiple times.

[0089] In the embodiment, the attribute maintainer is a model for maintaining attribute fields in the public data table. The attribute fields include attributes, states, etc. The attribute fields can be original fields in the public data table or fields additionally established in the public data table. If there are many attribute fields in the public data table, an attribute table can be separately established for storage. For example, when the states of historical data need to be recorded, an independent attribute table can be used for storage by period.

[0090] In the embodiment, the read-only accesser is a model that can only read data in the public data table. The read-only access model only has data reading permission and does not affect the data itself and other models.

[0091] Further, in some embodiments, the configuration of the data owner on the public data table specifically includes:

[0092] configuring basic information in the public data table;

[0093] configuring attribute fields maintained by the attribute maintainer;

[0094] giving data modification permission of the public data table, the data modification permission including permission for modifying data in the public data table;

[0095] configuring the number of times of calling the data owner as 1.

[0096] In the embodiment, referring to Figure 3 , the configuration content of the data owner includes: 1) configuring the table name, the description and other basic information of the public data table, and the data owner can selectively record whether the data in the public data table is updated; 2) configuring the attribute fields maintained by the attribute maintainer; 3) giving data modification permission of the public data table, the data modification permission including permission for modifying data in the public data table; the data owner has the unique data modification permission of the public data table, that is, the data in the public data table can only be modified by the data owner; 4) configuring the number of times of calling the data owner as 1; in order to ensure that the data in the public data table can only be modified once and avoid errors caused by multiple modifications of the data, the data owner is configured to be called only once, and when checking the number of times of calling the data owner, the calling situation of the upstream model can be checked layer by layer until the top model, and whether the data owner is called multiple times is judged.

[0097] Further, in some embodiments, the configuration of the attribute maintainer on the public data table specifically includes:

[0098] configuring the public data table to be maintained;

[0099] configuring the attribute fields to be maintained;

[0100] Check if the maintenance frequency of the attribute field matches the upstream data.

[0101] In the embodiment, referring to Figure 3 , the configuration content of the attribute maintainer includes: 1) configuring the common data table that needs to maintain the attribute, and the attribute maintainer can also selectively record whether the attribute in the common data table is updated; 2) configuring the attribute field that needs to be maintained; 3) checking if the maintenance frequency of the attribute field matches the upstream data, for example, the method can set the attribute timing maintenance frequency, when the maintenance frequency of the attribute field is higher than the update frequency of the upstream data, then the attribute field cannot obtain the latest upstream data when maintaining. Therefore, when checking if the maintenance frequency of the attribute field matches the upstream data, the method requires that the maintenance frequency of the attribute field is less than or equal to the update frequency of the upstream data, so as to ensure that the attribute field can obtain the latest upstream data when maintaining.

[0102] Further, in some embodiments, the permission of the read-only access to the common data table specifically includes:

[0103] Configuring the common data table to be read;

[0104] Checking if the data among the common data tables is synchronized.

[0105] In the embodiment, referring to Figure 3 , the configuration content of the read-only access includes: 1) configuring the common data table to be read, so that when the data owner and the attribute maintainer modify or maintain the common data table, the influence range of the modification or maintenance on other models can be clearly queried; 2) checking if the data among the common data tables is synchronized, when the read-only access needs to read multiple common data tables, it is necessary to judge whether the data update frequencies of the common data tables are synchronized, for example, the read-only access needs to judge whether the data of multiple common data tables has been updated, whether they are all the latest data, etc. The method can also display all the common data tables that are accessed, for judging whether the data in the common data tables is updated.

[0106] Further, in some embodiments, the bank risk control model is divided into real-time model, post-model and quasi-real-time model according to time effectiveness;

[0107] The data update mode of the real-time model is: real-time receiving data for updating;

[0108] The data update mode of the post-model is: batch importing data for updating in a long term;

[0109] The data update mode of the quasi-real-time model is: updating through data synchronization in a short term.

[0110] In the embodiment, the method divides the bank risk control model into three categories according to time effectiveness: real-time model, post-event model and quasi-real-time model. The real-time model updates data by receiving messages in real time, and the real-time model can be triggered by messages. The model has high requirements for time effectiveness and triggering frequency, so the calculation time (generally seconds or milliseconds) should be grasped when designing the real-time model, the algorithm is simple, and it does not need data from too many data sources.

[0111] The post-event model imports data in batches for updating in a fixed long time, for example, the post-event model imports the data of the business that has ended in batches, for example, some post-event models of the bank risk control model import the data after the day ends in batches. The post-event model has low time effectiveness, complex algorithm and large quantity. Since the post-event model needs to comprehensively analyze data from multiple data sources, the post-event model can read the update time points of each data source to determine whether the data required by the model is ready for calculation when all the data is ready. The data update frequency of the post-event model can be confirmed according to actual conditions, and the time is generally long, which can be one day, one month, etc.

[0112] The quasi-real-time model updates data through data synchronization in a short period of time. The data update frequency of the quasi-real-time model can be confirmed according to actual conditions, and the time is generally short, which can be several minutes, etc. For example, some quasi-real-time models perform batch data synchronization every 10 minutes. The data update frequency of the quasi-real-time model is high and has a certain complexity. In order to achieve good time effectiveness, the quasi-real-time model can reduce the public data level compared with the post-event model, so as to avoid the problem that when a public data level is wrong, it will affect the calculation of the subsequent public data level, resulting in poor real-time performance. In addition, in order to achieve good time effectiveness, the quasi-real-time model can establish a model private table for storing data, so that the quasi-real-time model can use the data in the model private table as much as possible when calculating, avoid calling other model processed data, and ensure the data time point uniformity. For quasi-real-time models with high precision requirements, complex algorithms and multiple data processing, the post-event model execution method can also be used when the system resources allow.

[0113] Further, the calling rule of the post-event model includes:

[0114] Obtaining data to be called;

[0115] Recording the preparation completion information of the data;

[0116] When it is judged according to the preparation completion information that the data is not ready, continue to wait for the preparation of the data;

[0117] When it is judged according to the preparation completion information that the data is ready, read all the data.

[0118] In the embodiment, refer toFigure 4 The method can configure the real-time model and the quasi-real-time model to have unique data write permissions and data modification permissions of the public data table, so that as long as the original data in the public data table is correct, the real-time model and the quasi-real-time model are not easy to make mistakes in the later calculation process, and the problem of poor real-time performance caused by a large number of calls to the processed data of other models is avoided. In order to ensure the uniformity of the data time point in the calling process, the following method is adopted: assuming that the post-event model needs two data sources, the data source can notify the post-event model after the data is prepared, and the post-event model records the preparation completion information of each data source. If the data preparation of the post-event model is not complete, continue to wait for the data source to prepare the data. If the data preparation of the post-event model is complete, read all the data for calculation, for example, when the data of the two data sources is prepared, update the public data table, and record the update state of the data in the public data table. After the post-event model is executed, it can be used as a data source for subsequent model calling.

[0119] The following examples further illustrate the configuration method and execution method of the bank risk control model in the method.

[0120] Referring to Figure 5 , assuming that the bank risk control model includes: bank risk control model a is used to alarm when a customer has more than 10 transfer transactions per day; bank risk control model b is used to screen customers whose transfer amount is more than 50% of the balance in a month. These two models need to call the public data table "transaction flow table" for daily update; add the daily debit and credit amount of the customer on the "transaction flow table" to form the "customer daily occurrence table"; add the daily amount on the "customer daily occurrence table" to form the "customer monthly occurrence table". Bank risk control model a finds the customers with more than 10 debit transactions on the "customer daily occurrence table" to perform early warning. Bank risk control model b compares the "customer monthly occurrence" with the last month's balance to screen.

[0121] The above two bank risk control models can also be split into sub-models in the following table, and the permission configuration method of each sub-model is shown in the following table:

[0122]

[0123] From the above table, it can be seen that each sub-model has unique write permission and attribute maintenance permission for each data, for example, the source data extraction sub-model has write permission for the transaction stream table, and the customer daily occurrence amount summary model has read permission for the transaction stream table and write permission for the daily summary table. This can better ensure data consistency, and the data update frequency of each sub-model matches its upstream data. This method can also automatically display the upper relationship of the model (the called model of the model) when setting the model, so that the calling relationship of the model and the public data table can be intuitively seen, and the permissions of each sub-model can be automatically judged, avoiding data confusion.

[0124] In addition, for the use scenario of Figure 1 , see Figure 6 , the bank risk control model inter-data sharing method records the converted renminbi of the model mdX and the model mdY in different fields, so that the model mdZ can call the correct field and read the correct converted renminbi.

[0125] Embodiment two:

[0126] A bank risk control model inter-data sharing system, comprising:

[0127] The configuration unit is configured to configure the public data table according to the bank risk control model when the bank risk control model registration needs to operate the public data table, and configure the parameters of the bank risk control model; the public data table includes data shared between the bank risk control models;

[0128] The execution unit is configured to read and write data in the public data table according to the configuration of the public data table when the bank risk control model is executed, and call the data in the public data table according to the calling rule.

[0129] Further, the bank risk control model is divided into a data owner, an attribute maintainer and a read-only access according to the relationship with the public data table;

[0130] The data owner is the model to which the data in the public data table belongs;

[0131] The attribute maintainer is the model that maintains the attribute field in the public data table;

[0132] The read-only access is the model that can only read the data in the public data table;

[0133] The data owner configures the public data table, specifically including:

[0134] Configuring the basic information in the public data table;

[0135] Configuring the attribute field for the attribute maintainer to maintain;

[0136] A data modification permission of the public data table is given, and the data modification permission comprises a permission of modifying data in the public data table;

[0137] The configuration data owner is called once;

[0138] The attribute maintainer configures the public data table, and the configuration specifically comprises:

[0139] The public data table to be maintained is configured;

[0140] The attribute field to be maintained is configured;

[0141] It is checked whether the maintenance frequency of the attribute field matches upstream data;

[0142] The read-only accesser configures the public data table, and the configuration specifically comprises:

[0143] The public data table to be read is configured;

[0144] It is checked whether data among the public data tables is synchronized.

[0145] Further, the bank risk control model is divided into real-time model, post-event model and quasi-real-time model according to time effectiveness;

[0146] The data updating mode of the real-time model is to update in real time by receiving data;

[0147] The data updating mode of the post-event model is to update by importing data in batches in a long term;

[0148] The data updating mode of the quasi-real-time model is to update by synchronizing data in a short term;

[0149] The calling rule of the post-event model comprises:

[0150] Data to be called is obtained;

[0151] Preparation completion information of the data is recorded;

[0152] When it is judged according to the preparation completion information that the data is not ready, the preparation of the data is continued to be waited for;

[0153] When it is judged according to the preparation completion information that the data is ready, all the data is read.

[0154] The system provided by the embodiment of the present application is described briefly, and the part not mentioned in the embodiment can refer to the corresponding content in the foregoing embodiment.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should be covered in the scope of the claims and the description of the present application.

Claims

1. A method for data sharing among bank risk control models, characterized in that, include: When a bank risk control model that needs to operate on a public data table is registered, the public data table is configured according to the bank risk control model, as are the parameters of the bank risk control model; the public data table includes data shared among bank risk control models; When the bank risk control model is executed, it reads and writes data in the public data table according to the configuration of the public data table, and calls data in the public data table according to the calling rules; The bank's risk control model is divided into three categories based on its relationship with the public data table: data owner, attribute maintainer, and read-only visitor. The data owner is the model to which the data in the public data table belongs; The attribute maintainer is the model that maintains the attribute fields in the public data table; The read-only visitor is a model that can only read data from the public data table; The specific configurations that a data owner makes for a public data table include: Configure the basic information in the public data table; Configure the attribute fields to be maintained by the attribute maintainer; Grant data modification permissions to the public data table, including the permission to modify data in the public data table; The data owner is configured to be invoked only once. The configuration of public data tables by the attribute maintainer specifically includes: Configure the public data tables that need to be maintained; Configure the attribute fields that need to be maintained; Check whether the maintenance frequency of the attribute field matches the upstream data; Its key feature is that the configuration of the public data table by a read-only access user specifically includes: Configure the reading of the public data table; Check whether the data in each of the aforementioned public data tables is synchronized.

2. The data sharing method among bank risk control models according to claim 1, characterized in that, The bank risk control models are classified according to their timeliness into: real-time models, post-event models, and near-real-time models. The real-time model is updated by receiving data in real time. The data update method for the post-event model is: to import data in batches and update it over a long period of time; The data update method of the near real-time model is to update it through data synchronization in the short term.

3. The data sharing method between bank risk control models according to claim 2, characterized in that, The rules for invoking the post-event model include: Retrieve the data that needs to be called; Record the completion information of the data preparation; If the data is not fully prepared based on the preparation completion information, continue to wait for the data to be prepared. When the data is deemed ready based on the preparation completion information, all the data is read.

4. A data sharing system between bank risk control models, characterized in that, include: Configuration unit: used to configure the public data table according to the bank risk control model when a bank risk control model that needs to operate on the public data table is registered, and to configure the parameters of the bank risk control model; the public data table includes data shared among bank risk control models; Execution Unit: Used when the bank risk control model is executed to read and write data in the public data table according to the configuration of the public data table, and to call data in the public data table according to the calling rules; The bank's risk control model is categorized according to its relationship with the public data table into: data owner, attribute maintainer, and read-only accessor. The data owner is the model to which the data in the public data table belongs; The attribute maintainer is the model that maintains the attribute fields in the public data table; The read-only visitor is a model that can only read data from the public data table; The specific configurations that a data owner makes for a public data table include: Configure the basic information in the public data table; Configure the attribute fields to be maintained by the attribute maintainer; Grant data modification permissions to the public data table, including the permission to modify data in the public data table; The data owner is configured to be invoked only once. The configuration of the public data table by the attribute maintainer specifically includes: Configure the public data tables that need to be maintained; Configure the attribute fields that need to be maintained; Check whether the maintenance frequency of the attribute field matches the upstream data; The configuration of the public data table by the read-only access user specifically includes: Configure the reading of the public data table; Check whether the data in each of the aforementioned public data tables is synchronized.

5. The data sharing system between bank risk control models according to claim 4, characterized in that, The bank risk control models are classified according to their timeliness into: real-time models, post-event models, and near-real-time models. The real-time model is updated by receiving data in real time. The data update method for the post-event model is: to import data in batches and update it over a long period of time; The data update method of the near real-time model is: to update through data synchronization in the short term; The rules for invoking the post-event model include: Retrieve the data that needs to be called; Record the completion information of the data preparation; If the data is not fully prepared based on the preparation completion information, continue to wait for the data to be prepared. When the data is deemed ready based on the preparation completion information, all the data is read.

Images