CN119006137A

CN119006137A - Risk decision method, system, electronic equipment and storage medium

Info

Publication number: CN119006137A
Application number: CN202310546538.1A
Authority: CN
Inventors: 冯宏轩; 卢秋明; 陈�光; 刘志浩
Original assignee: Bairong Yunchuang Technology Co ltd
Current assignee: Bairong Yunchuang Technology Co ltd
Priority date: 2023-05-16
Filing date: 2023-05-16
Publication date: 2024-11-22
Also published as: WO2024234855A1

Abstract

The present application discloses a risk decision-making system with a feature derivation function, including a feature derivation platform, a model center subsystem, a business product management subsystem, and a risk report generation subsystem. The feature derivation platform is configured to extract basic variables based on incoming data, third-party data, and internal data, generate derived variables based on the basic variables, and generate data sets based on the basic variables and the derived variables; the model center subsystem is configured to configure a risk control model according to user selection; the business product management subsystem is configured to create a business product, associate the data set with the risk control model, and start the business product; the risk report generation subsystem is configured to generate a risk report based on the operating results of the business product. The present application can quickly implement risk decisions for different business products and reduce the need to repeatedly build a risk control system. Corresponding methods, electronic devices, and computer-readable storage media are also provided.

Description

Risk decision method, system, electronic equipment and storage medium

Technical Field

The application relates to the technical field of computer networks, in particular to a risk decision system with a feature deriving function. In addition, the application also relates to a related method, an electronic device and a computer readable storage medium.

Background

With the rapid development of internet financial services, fraud and crimes aiming at financial institutions are increased, diversified and technological characteristics are presented, a great risk is caused to the financial institutions, and meanwhile, the policies of the national supervision departments clearly specify that the banks must have independent and independent full-flow wind control management capability when the internet financial services are developed, so that a matched full-flow wind control system is required to be established while the internet financial services are developed, the financial institutions need to make credit decisions more quickly, and credit approval timeliness is improved. However, the approval timeliness is improved, and meanwhile, the credit approval risk is also improved, and a more perfect automatic approval process is required, so that a risk decision system with a feature derivative function is required to be constructed, a credit approval system is perfected, customer credit approval is efficiently and accurately carried out, and the credit risk assessment and control capability is improved. However, the existing risk decision system with the feature derivative function is often a system with a fixed model, different systems are required to be built aiming at different business products, the cost is high, and the efficiency is low.

The description of the background art is only for the purpose of facilitating an understanding of the relevant art and is not to be taken as an admission of prior art.

Disclosure of Invention

Therefore, the invention aims to provide the risk decision system and the method with the feature derivation function, which can rapidly realize the construction of the wind control model of the business product in a collocation mode, improve the wind control decision capability of the business product, improve the research and development efficiency and rapidly respond to market change.

Specifically, the invention provides a risk decision system with a feature deriving function, which comprises a feature deriving platform, a model center subsystem, a business product management subsystem and a risk report generating subsystem, wherein,

The feature deriving platform is configured to extract basic variables according to the feed data, the three-party data and the internal data, generate derived variables according to the basic variables, and generate a data set according to the basic variables and the derived variables;

The model center subsystem is configured to configure a wind control model according to user selection;

The business product management subsystem is configured to create a business product, associate the business product with a corresponding data set in the feature derivation platform according to user selection, create a wind control model associated with the business product, set the data set corresponding to the business product to a variable library of the wind control model, invoke a model center subsystem according to user selection, configure the wind control model associated with the business product, and enable the business product to operate;

the risk report generation subsystem is configured to generate a risk report according to the operation result of the business product.

In some embodiments of the present invention, the business product management subsystem is further configured to perform a simulation test on the wind control model, and generate a simulation test result.

In some embodiments of the present invention, the incoming data includes a transaction message, and the feature derivation platform is specifically configured to automatically parse the transaction message to obtain a basic variable in the transaction message; and generating derivative variables according to the configuration rules of the derivative variables based on the basic variables, wherein the transaction messages comprise personal second-generation credit message, enterprise credit message and third-party message.

In some embodiments of the invention, the business product management subsystem is configured to export the business product according to user selection and export a dataset and an air control model associated with the business product.

In some embodiments of the invention, the business product management subsystem is configured to import an external business product according to user selection and import a dataset and a wind control model associated with the external business product.

Preferably, in an embodiment of the present invention, when the model central subsystem configures the wind control model associated with the service product, the model central subsystem is specifically configured to configure the wind control model by using a policy management component, where the policy management component includes at least one of a decision set management component, a decision table management component, a decision tree management component, a score card management component, a machine learning model management component, and a decision stream management component.

In some embodiments of the invention, the risk report includes the results of the transaction, the feed data of the transaction, the three-way data, the internal data, and the rules of the hit.

In some embodiments of the invention, the system further comprises a model optimization subsystem configured to optimize the wind control model according to an optimization mode including at least one of an AB test mode, a champion challenger mode, a rule backtracking mode.

In some embodiments of the present invention, there is further provided a risk decision method applied to a risk decision system with feature derivation function, where the risk decision system with feature derivation function includes a feature derivation platform, a model center subsystem, a business product management subsystem, and a risk report generation subsystem, and the method includes:

The feature derivation platform extracts basic variables according to the feed data, the three-party data and the internal data, generates derived variables according to the basic variables, and generates a data set according to the basic variables and the derived variables;

the model center subsystem configures a wind control model according to user selection;

The business product management subsystem creates a business product, associates the business product with a corresponding data set in the feature derivative platform according to user selection, creates a wind control model associated with the business product, sets the data set corresponding to the business product to a variable library of the wind control model, invokes a model center subsystem according to user selection, configures the wind control model associated with the business product, and enables the business product to operate;

And the risk report generating subsystem generates a risk report according to the operation result of the business product.

In some embodiments of the invention, the feature-derived risk decision system further comprises a model optimization subsystem, the method further comprising:

The model optimization subsystem optimizes the wind control model according to an optimization mode, wherein the optimization mode comprises at least one of an AB test mode, a champion challenger mode and a rule backtracking mode.

In some embodiments of the present invention, a computer readable storage medium is provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the risk decision method of any of the embodiments of the present invention.

In some embodiments of the present invention, there is provided an electronic device including: a processor and a memory storing a computer program, the processor being configured to perform any of the risk decision methods of the embodiments of the present invention when the computer program is run.

The embodiment of the invention provides a risk decision system and a risk decision method with a feature derivative function, wherein a business product is built in the system, an incoming data and an air control model are associated, the corresponding air control model is selected for configuration, incoming data is associated to the air control model, the business product is started to operate, a risk report is generated according to the operation result of the business product, a configurable risk decision system with the feature derivative function is realized, the response capability of risk decisions for different business products is improved, the risk decisions can be realized rapidly for different business products, and the requirement of repeatedly building the air control system is reduced. Basic variables are extracted through incoming data, based on derivative rules, living derived variables enrich the data types of wind control model decisions, and wind control model decision results are more accurate. The decision system in the embodiment of the invention can support the whole industry credit wind control business, and is locally deployed, so that unified integration, operation and management of rules, data and models are realized, friendly operation pages are provided, and wind control strategy configuration type operation is provided for wind control personnel. The system can interact with the business system in an interface mode to provide decision support for the credit approval system. Through the integrated application of internal and external credit data and the flexible setting of scene-based wind control strategies and the combination of various auxiliary technologies, the wind control decision-making capability of the full life cycle of credit business in the fields of consumer finance, credit card, benefits, public and the like is comprehensively improved, meanwhile, the accumulation, multiplexing and management of the full-line credit decision-making capability are realized, the demand research and development efficiency is continuously improved, and the market change is responded quickly.

Additional optional features and technical effects of embodiments of the invention are described in part below and in part will be apparent from reading the disclosure herein.

Drawings

Embodiments of the present invention will hereinafter be described in conjunction with the appended drawings, wherein like or similar reference numerals denote like or similar elements, and wherein:

FIG. 1 shows a system architecture diagram of a risk decision system with feature derivation functions;

FIG. 2 illustrates a user operational flow diagram of a risk decision system with feature derivation functionality;

FIG. 3 shows a flow diagram of a risk decision method;

FIG. 4 shows another flow diagram of a risk decision method;

Fig. 5 shows an exemplary structural schematic of an electronic device capable of implementing the method according to an embodiment of the invention.

Detailed Description

The present invention will be described in further detail with reference to the following detailed description and the accompanying drawings, in order to make the objects, technical solutions and advantages of the present invention more apparent. The exemplary embodiments of the present invention and the descriptions thereof are used herein to explain the present invention, but are not intended to limit the invention.

In the present embodiment, "KEEPALIVE" is a mechanism that can detect dead connections; "nginx" is a high performance HTTP and reverse proxy web server; "Vue" is a front end frame;

"ES6" is a front-end technology; "SASS" is a cascading style sheet language; "Echarts" is a library of data visualization charts; "Element-UI" is a library of user interaction components; "Restful" is a design style and development style of web applications; "Kafka" is an open source stream processing platform; "RocketMQ" is a messaging system; "MySQL" is a database; "Oracle" is a relational database, and "SQLServer" is a database management system; "PostgreSQL" is a relational database management system; "NoSQL" is a non-relational database; "Redis" is a key value storage system; "MongoDB" is a database based on distributed file storage; "Jackrabbit" is a document-oriented database; "rpc" is a remote procedure call technique; "Future" is a model of asynchronous task processing.

As shown in fig. 1, an embodiment of the present invention provides a risk decision system 100 with feature derived functionality, including a feature derived platform 110, a model center subsystem 120, a business product management subsystem 130, and a risk report generation subsystem 140, wherein,

The feature derivation platform 110 is configured to extract basic variables according to the feed data, the three-party data and the internal data, generate derived variables according to the basic variables, and generate a data set according to the basic variables and the derived variables;

The model center subsystem 120 is configured to configure a wind control model according to user selection;

The service product management subsystem 130 is configured to create a service product, associate the service product with a corresponding data set in the feature derivation platform according to user selection, create a wind control model associated with the service product, set the data set corresponding to the service product to a variable library of the wind control model, invoke a model center subsystem 120 according to user selection, configure the wind control model associated with the service product, and enable the service product to operate;

The risk report generating subsystem 140 is configured to generate a risk report from the operational results of the business product.

The system in the embodiment of the invention can create the business product, can correlate the data set and the wind control model, loads the data set into the variable library of the wind control model, enables the business product to operate after the user configures the wind control model through the visualized interface, and automatically generates the risk report based on the operation structure of the business product, thereby improving the efficiency of building the wind control system and being capable of rapidly responding to different business products. The business product in the embodiment of the invention can be credit product, credit card product and the like.

The system in some embodiments of the invention can support the whole life cycle of products such as consumer finance, credit card, benefits, public, and the like, and has good support for strategies and models before, during and after credit. Meanwhile, the supported business products are also various, and the system is designed from two dimensions of life cycle and business product types, so that the deployment of most business product strategies and models can be covered. By flexibly configuring the relevant rules in the decision engine, risk management can be performed throughout the credit product lifecycle. Before lending, the method can realize application anti-fraud, equipment anti-fraud, credit evaluation, product pricing, real-time credit decision and the like; in the lending, the dynamic early warning of stock clients can be realized; after the credit, the collection prompting strategy can be formulated.

In some embodiments of the present invention, the service product management subsystem 130 is further configured to perform a simulation test on the wind control model to generate a simulation test result. The effectiveness of the wind control model can be rapidly verified, and the construction efficiency is improved.

In some embodiments of the present invention, the incoming data includes a transaction message, and the feature derivation platform 110 is specifically configured to automatically parse the transaction message to obtain a basic variable in the transaction message; and generating derivative variables according to the configuration rules of the derivative variables based on the basic variables, wherein the transaction messages comprise personal second-generation credit message, enterprise credit message and third-party message.

The use requirement of a risk department on pedestrian variables is met by analyzing and deriving the variables in reports such as second-generation pedestrian credit and the like. And the online management of the characteristic variables is provided, and the real-time receiving, reading, calculating and service are realized. The method comprises the steps of realizing unified management of characteristic variables, formulating the variables, deriving logic configuration, receiving, reading and processing the variables in real time and the like, more efficiently adapting to the implementation of second-generation credit investigation, and providing solid basic data service for a decision engine and a model training platform.

In some embodiments of the invention, the feature derivation process includes:

) And an analyst translates variables required by the wind control model into specific logic according to requirements provided by service personnel.

) Derivative variable configuration. And entering a variable scheme new page through a new scheme button. The newly added variable scheme is divided into 4 steps, namely 4 pages, and the steps are respectively as follows: basic information, message import, basic variable correction and characteristic variable configuration. Basic information of the basic variable page filling scheme comprises scheme names, data sources and the like. The message import page realizes the sample message import corresponding to the scheme and prepares for subsequent analysis of basic variables and configuration variables. The basic variable correction page is used for checking basic variables after sample message analysis and supporting page modification. The derived variable configuration page configures derived variables according to the basic message and supports testing, start-stop and version management functions.

) Variable release. After the configuration of the derived variables is completed, the configuration page of the variables is started, and after the scheme is started, the scheme and the variables are in a release state. The user can import the page in the variable, import the message data in real time, and can view the derived variable result on the page

) And (5) feature report. A user can manage the page through the interface and configure a corresponding interface, a third party system can push the message to the feature deriving platform 110 through the interface form, the feature deriving platform 110 performs real-time calculation, and the synchronous interface can return derived variable settlement results in real time to form a data set.

In the embodiment of the invention, the feature deriving platform 110 is internally provided with common feature deriving logic, is internally provided with more than 1.6w of personal/enterprise second-generation credit deriving variables, and can respectively list the variables of 20 before validity in combination with scenes; the built-in derivative variable template system is provided with more than 80 functions, so that the derivative requirement is met, the uploading of the custom function is supported, and the requirement of a user is met to the maximum extent. The derivative platform supports a developer to develop a custom function by using java codes, and the custom function is imported into the system by a page importing mode. The development efficiency can be improved by using the derivative variable configuration; feature variables are configured in a plurality of ways, including page configuration expression, editing using SQL statements, JS codes, python codes, and the like.

For example, a basic variable { X1, X2, X3} is extracted from the message data, for example, { X1, X2, … xn } is calculated for the number of loan repayment overdue days { X1, X2, … xn } in a certain loan product repayment sequence, a derivative variable { X1, X2, X3} is generated based on a derivative rule, for example, a rule for summing up the number of data, and a rule for summing up the variance of the data, and a data set DataX= { X1, X2, … xn, X1, X2, X3} is generated based on the basic variable and the derivative variable. After the loan product is created, the business product management subsystem 130 is associated with the corresponding dataset DataX and the air control model M. The wind control model M is configured by invoking the model center subsystem 120.

In some embodiments of the present invention, when the model central subsystem 120 configures the wind control model associated with the service product, the model central subsystem 120 is specifically configured to configure the wind control model with a policy management component, where the policy management component includes at least one of a decision set management component, a decision table management component, a decision tree management component, a score card management component, a machine learning model management component, and a decision stream management component.

In some embodiments of the present invention, when the decision set management component is configured, the complex rule customization can be completed by clicking the mouse; when the configuration is carried out through the decision table management component, a specific complex type rule is defined in a decision table form, and the condition or action configuration can be completed through clicking a mouse, so that the configuration method is suitable for business personnel to be familiar with; when the configuration is carried out through the decision tree management component, the service rules are expressed in a visual and tree-shaped structure; when the configuration is carried out through the grading card management component, the graphical grading card editor is matched with the grading card in the business system to the maximum extent; when the complex machine learning model file deployment is carried out through the machine learning model management component, the complex machine learning model file deployment is supported; when the decision flow management component is used for configuration, the arrangement execution of a plurality of complex business rules is realized, the decision flow is completed by dragging a mouse based on the browser realization decision, a rule set, a decision table, a grading card and a decision tree can be bound on a flow node, the rule arrangement is realized, the flow nodes of various industry types such as routing judgment, serial, parallel and the like are provided in the flow, and the rule arrangement of various complex businesses is met to the maximum extent.

Preferably, in some embodiments of the present invention, in order to handle concurrent product creation tasks, the service product management subsystem 130 further has an import-export function, and the service product management subsystem 130 is configured to export the service product according to user selection and export a dataset and an air control model associated with the service product. The business product management subsystem 130 may import the created business product into other system executions. After test verification is carried out in the system, the system is conveniently imported into a production environment for execution, and the deployment efficiency is improved.

In some embodiments of the present invention, the business product management subsystem 130 is configured to import external business products according to user selections and import datasets and wind control models associated with the external business products.

In some embodiments of the present invention, the risk report generating subsystem 140 may be configured to automatically generate a risk report according to an operation result of a service product, where the risk report includes a result of a transaction, feed data of the transaction, three-party data, internal data, hit rules, and the like, and may also generate a corresponding chart, PDF file, and the like.

In some embodiments of the present invention, a plurality of ways are provided for model optimization, the system further comprising a model optimization subsystem 150, the model optimization subsystem 150 being configured to optimize the wind control model according to an optimization mode, the optimization mode comprising at least one of the following optimization modes.

1) AB test mode, the optimization process is: the model A is a model used by the current business product, the model B is a new model, one transaction data is used for executing the models A and B at the same time, normal business cannot be influenced, and the execution results of the models A and B can be compared to find a better model.

2) Champion challenger mode, the optimization process is as follows: the champion is a model used by the current business product, the challenger is a new model, a plurality of challengers can be configured, then an execution proportion is set, the model which is supposed to be executed can be calculated according to the designed proportion for each input piece, and one input piece only executes one model at the same time.

3) Rule backtracking mode, the optimization process is as follows: running the new model with the previous feed data.

The system in the embodiment of the invention can support the creation of consumer finance, credit card, benefits, public products and the like in the whole life cycle, and can well support strategies and models before credit, during credit and after credit. Meanwhile, the supported business products are also various, and the system is designed from two dimensions of life cycle and business product types, so that the deployment of most business product strategies and models can be covered. By flexibly configuring the relevant rules in the decision engine, risk management can be performed throughout the credit product lifecycle. Before lending, the method can realize application anti-fraud, equipment anti-fraud, credit evaluation, product pricing, real-time credit decision and the like; in the lending, the dynamic early warning of stock clients can be realized; after the credit, the collection prompting strategy can be formulated.

The system in the embodiment of the invention supports flexible deployment, updating, maintenance and management of various scoring models and strategies, reduces uncontrollable human factors of the manual credit flow, gradually improves the standardized and differentiated management capability of risk models and strategies before, during and after credit business credit, and improves the automation degree of model and strategy operation. The detailed functions are described as follows:

1) And supporting risk policy model management, including risk policy model definition, maintenance, release and the like, and performing policy model iteration and completing release in real time.

2) Rich policy management components, such as rules/rule sets, decision tables, decision trees, scoring models, data computations, decision flows, etc., are supported, and corresponding validity period, policy priority, etc., management is supported.

① Rules: the decision logic is organized in a logical expression statement, such as "… if …".

② Rule set: the formed rule set is organized based on the unified input parameters and the processing logic scope.

③ Decision tree: using tree branch pattern to calculate: the processing calculation is carried out by inputting variables or application form data elements, parameters required by subsequent rule judgment or risk calculation are obtained, and the parameters can be used in other decision components.

④ Decision table: the decision logic is organized using a tabular graphical approach.

⑤ Decision flow: the plurality of decision components are arranged in a manner to implement overall decision logic.

⑥ Scoring model: and supporting a scoring card model, and evaluating and quantifying credit risks according to a scoring card mode and obtaining specific scores.

⑦ Champion/challenger mode: the champion/challenger implementation mode can be formulated, a plurality of challenger strategies are designated, the system performs statistical evaluation on execution efficiency, operation effect and the like of the plurality of challenger strategies, and the optimal challenger strategy is upgraded into the champion strategy.

⑧ The graphical operation is supported, the attractive UI interface is provided, a user can model in a dragging mode, model creation can be completed without programming experience, and the decision logic is supported to be organized by dragging and selecting a data source.

3) Support selection algorithm, configure decision flow, configure parameters on interface, filter variables, simply process data, view algorithm execution process, etc.

4) Flexible and complex data structures are supported, including business object models based on XML schemas, or business object models expressed using hierarchically nested Java objects. For example, a complex hierarchical data structure of the pedestrian credit report may be implemented so that policy administrators may create derivative variables based on the detailed data of the pedestrian credit report.

5) And the authority management is supported, the authority management strategy can be flexibly formulated, and the rule range accessible to the roles is limited. Rights control may be performed according to a rule base directory/package. One user may be granted multiple roles.

6) And the system supports approval management of the changed strategy model, and the changed strategy model can be issued only after approval of the related authority user.

7) And the audit of operation logs of all rule bases is supported, and various detail queries are carried out, including which user logs in the rule base at what time and which operation is carried out.

Preferably, the system in the embodiment of the invention comprises:

1. Feature derivation platform: the sources of variables used in the model may originate from the feed of some other business system, or from the query of interfaces provided by the querying party and from querying some database, the subsystem providing unified integration and management of such data, being able to manage the field verification rules, set the feed fields, and maintain the data products. The feature derived platform in the embodiment of the invention can perform regular expression verification configuration on the entry fields, uniformly and intensively perform entry field maintenance, uniformly and intensively maintain internal and external data products, support multiple modes, and reduce later code development through XML file configuration.

2. And a business product management subsystem: creating different business products for different business scenes, associating the business products with the models of the in-process, in-process and out-process data and machine learning used by the scenes, automatically creating a model item and the business products for association, automatically creating selected data in the business products into a variable library of the model, and facilitating the model to directly use variable modeling; the business product can automatically generate an api file to provide service for the outside and provide a simulation test function; the business products provide the functions of importing and exporting, and import and export the selected business products, the internal and external machine learning models and the related models, so that the business products which are measured by the testing environment are exported and imported into the production environment when the production environment is online and can be directly used.

The business product management subsystem in the embodiment of the invention associates the business products with the used wind control model and the data set (data products), can perform maintenance and test, can transplant the business products to other environments, can also enable or disable the business products, and has a safety control mechanism.

3. Model center subsystem: the mode of providing page clicking configures the model, supports multiple models, the mode of providing decision flow arranges other models into one flow, and the function of providing simulation test verifies the accuracy of the model. When the model center subsystem configures the wind control model associated with the service product, the model center subsystem is specifically configured to configure the wind control model by adopting a policy management component, wherein the policy management component comprises at least one of a decision set management component, a decision table management component, a decision tree management component, a scoring card management component, a machine learning model management component and a decision flow management component.

In the embodiment of the invention, when the decision set management component is used for configuration, the complex rule customization can be completed by clicking a mouse; when the configuration is carried out through the decision table management component, a specific complex type rule is defined in a decision table form, and the condition or action configuration can be completed through clicking a mouse, so that the configuration method is suitable for business personnel to be familiar with; when the configuration is carried out through the decision tree management component, the service rules are expressed in a visual and tree-shaped structure; when the configuration is carried out through the grading card management component, the graphical grading card editor is matched with the grading card in the business system to the maximum extent; when the complex machine learning model file deployment is carried out through the machine learning model management component, the complex machine learning model file deployment is supported; when the decision flow management component is used for configuration, the arrangement execution of a plurality of complex business rules is realized, the decision flow is completed by dragging a mouse based on the browser realization decision, a rule set, a decision table, a grading card and a decision tree can be bound on a flow node, the rule arrangement is realized, the flow nodes of various industry types such as routing judgment, serial, parallel and the like are provided in the flow, and the rule arrangement of various complex businesses is met to the maximum extent.

4. Risk report generation subsystem: for each transaction, providing detail data of risk report query transaction, including the result of the transaction, the entry data of the transaction, the queried three-party data and internal data, the details of hit rules, etc., and may generate a PDF file download.

5. Model optimization subsystem: various ways are provided for model optimization.

AB test: the model A is a model used by the current business product, the model B is a new model, one transaction data is used for executing the models A and B at the same time, normal business cannot be influenced, and the execution results of the models A and B can be compared to find a better model.

Champion challenger: the champion is a model used by the current business product, the challenger is a new model, a plurality of challengers can be configured, then an execution proportion is set, one input part can calculate the model which should be executed according to the designed proportion, and one input part can execute only one model at the same time.

Rule backtracking: running the new model with the previous feed data.

Preferably, the system in the embodiment of the invention comprises three parts: load layer, service layer, middleware.

And load balancing in the load layer adopts KEEPALIVE and nginx to distribute services, so that high availability of the application is ensured.

The service layer adopts a mainstream MVC architecture and an efficient data caching tool, performs WEB end display by matching with front-end technologies Vue, ES6, SASS, echarts, element-UI and the like, and simultaneously provides a Restful service interface by adopting an SAAS development mode.

The message queue is Kafka or RocketMQ for realizing asynchronous transaction, which can be selected according to the actual situation of the client, and ensures the service decoupling of the decision engine and the service system and the stability and reliability of the service.

The relational database decision engine is compatible with MySQL and Oracle by default, while supporting adaptation to SQLServer and PostgreSQL.

NoSQL selects Redis, can carry on the deployment mode of cluster/sentry/master-slave/stand-alone according to the business requirement, offer the cache for business transaction of the business, guarantee the fast response of the transaction. And simultaneously, mongoDB, jackrabbit is selected for file storage.

The user-defined rpc remote procedure call protocol is adopted for the internal services, so that the transmission efficiency between the services is improved, the service performance consumption is reduced, and the service management is facilitated.

The external service adopts a custom http call protocol, and mainly provides an external heterogeneous environment, browser interface call, app interface call, third party interface call and the like.

And performing task execution of a model countermeasure and model backtracking strategy by adopting a Future mode in asynchronous task processing to obtain a report difference comparison result, and adjusting rules to optimize the model.

As shown in fig. 2, when the system in the embodiment of the present invention is used from the viewpoint of the operation of the service personnel, the operation steps are as follows:

S210: the required feed fields, internal data and external data are configured into the system in the feature derivation platform.

S220: a business product is created to sort out the required incoming, internal and external data.

S230: clicking model configuration on the created service product, entering a menu of a model center, configuring a model in a page clicking mode, newly establishing a strategy package in model management, referencing the configured model file in the strategy package, and verifying the accuracy of the model through simulation test in the strategy package.

S240: and selecting a policy package built in model management on the service product, starting the service product, and calling the service product through an api file of the service product.

S250: the transaction details are viewed in a risk report.

The embodiment of the invention provides a risk decision system with a feature derivative function, a business product is built in the system, incoming data and an air control model are associated, the corresponding air control model is selected for configuration, the incoming data is associated to the air control model, the business product is started to operate, a risk report is generated according to the operation result of the business product, the configurable risk decision system with the feature derivative function is realized, the risk decision response capability for different business products is improved, the risk decision can be realized rapidly for different business products, and the requirement for repeatedly building the air control system is reduced. Basic variables are extracted through incoming data, based on derivative rules, living derived variables enrich the data types of wind control model decisions, and wind control model decision results are more accurate. The decision system in the embodiment of the invention can support the whole industry credit wind control business, and is locally deployed, so that unified integration, operation and management of rules, data and models are realized, friendly operation pages are provided, and wind control strategy configuration type operation is provided for wind control personnel. The system can interact with the business system in an interface mode to provide decision support for the credit approval system. Through the integrated application of internal and external credit data and the flexible setting of scene-based wind control strategies and the combination of various auxiliary technologies, the wind control decision-making capability of the full life cycle of credit business in the fields of consumer finance, credit card, benefits, public and the like is comprehensively improved, meanwhile, the accumulation, multiplexing and management of the full-line credit decision-making capability are realized, the demand research and development efficiency is continuously improved, and the market change is responded quickly.

As shown in fig. 3, the embodiment of the present invention further provides a risk decision method, applied to a risk decision system with feature derivation function, where the risk decision system with feature derivation function includes a feature derivation platform, a model center subsystem, a service product management subsystem, and a risk report generation subsystem, and the method includes:

s310: the feature derivation platform extracts basic variables according to the feed data, the three-party data and the internal data, generates derived variables according to the basic variables, and generates a data set according to the basic variables and the derived variables;

S320: the model center subsystem configures a wind control model according to user selection;

S330: the business product management subsystem creates a business product, associates the business product with a corresponding data set in the feature derivative platform according to user selection, creates a wind control model associated with the business product, sets the data set corresponding to the business product to a variable library of the wind control model, invokes a model center subsystem according to user selection, configures the wind control model associated with the business product, and enables the business product to operate;

S340: and the risk report generating subsystem generates a risk report according to the operation result of the business product.

In some embodiments of the present invention, the risk decision system with feature derivation function further includes a model optimization subsystem, as shown in fig. 4, the method includes:

s410: the feature derivation platform extracts basic variables according to the feed data, the three-party data and the internal data, generates derived variables according to the basic variables, and generates a data set according to the basic variables and the derived variables;

s420: the model center subsystem configures a wind control model according to user selection;

S430: the business product management subsystem creates a business product, associates the business product with a corresponding data set in the feature derivative platform according to user selection, creates a wind control model associated with the business product, sets the data set corresponding to the business product to a variable library of the wind control model, invokes a model center subsystem according to user selection, configures the wind control model associated with the business product, and enables the business product to operate;

s440: and the risk report generating subsystem generates a risk report according to the operation result of the business product.

S450: the model optimization subsystem optimizes the wind control model according to an optimization mode, wherein the optimization mode comprises at least one of an AB test mode, a champion challenger mode and a rule backtracking mode.

The operation steps and effects of each subsystem in the method of the embodiment of the present invention refer to the operation steps and effects of the corresponding subsystem in the above system, and are not described herein.

In an embodiment of the present invention, there is provided an electronic device including: a processor and a memory storing a computer program, the processor being configured to perform any of the risk decision methods of the embodiments of the present invention when the computer program is run.

Fig. 5 shows a schematic diagram of an electronic device 500 that may implement or implement embodiments of the present invention, and in some embodiments may include more or fewer electronic devices than shown. In some embodiments, it may be implemented with a single or multiple electronic devices. In some embodiments, implementation may be with cloud or distributed electronic devices.

As shown in fig. 5, the electronic device 500 includes a processor 501 that can perform various suitable operations and processes according to programs and/or data stored in a Read Only Memory (ROM) 502 or programs and/or data loaded from a storage portion 508 into a Random Access Memory (RAM) 503. Processor 501 may be a multi-core processor or may include multiple processors. In some embodiments, processor 501 may comprise a general-purpose main processor and one or more special coprocessors such as, for example, a Graphics Processor (GPU), a neural Network Processor (NPU), a Digital Signal Processor (DSP), etc. In the RAM503, various programs and data required for the operation of the electronic apparatus 500 are also stored. The processor 501, ROM502, and RAM503 are connected to each other by a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.

The processor and the memory are used together to execute a program stored in the memory, which when executed by a computer, enables the steps or functions of the risk decision method described in the above embodiments to be implemented.

The following components are connected to the I/O interface 505: an input section 506 including a keyboard, a mouse, a touch screen, and the like; an output portion 507 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker, and the like; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The drive 510 is also connected to the I/O interface 505 as needed. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as needed so that a computer program read therefrom is mounted into the storage section 508 as needed. Only some of the components are schematically shown in fig. 5, which does not mean that the computer system 500 includes only the components shown in fig. 5.

In some embodiments, the electronic device 1020 refers to a mobile terminal, including a mobile phone, a vehicle-mounted terminal, a smart television, etc., and the electronic device 500 further includes a display screen with a touch function, an externally-placed sound box, a gyroscope, a camera, a 4G/5G antenna, etc.

The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer or its associated components. The computer may be, for example, a mobile terminal, a smart phone, a personal computer, a laptop computer, a car-mounted human-computer interaction device, a personal digital assistant, a media player, a navigation device, a game console, a tablet, a wearable device, a smart television, an internet of things system, a smart home, an industrial computer, a server, or a combination thereof.

Although not shown, in an embodiment of the present invention, a storage medium is provided, the storage medium storing a computer program configured to, when executed, perform any of the risk decision methods of the embodiments of the present invention.

Storage media in embodiments of the invention include both permanent and non-permanent, removable and non-removable items that may be used to implement information storage by any method or technology. Examples of storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, read only compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium which can be used to store information that can be accessed by a computing device.

Methods, programs, systems, apparatus, etc. in accordance with embodiments of the invention may be implemented or realized in single or multiple networked computers, or in distributed computing environments. In the present description embodiments, tasks may be performed by remote processing devices that are linked through a communications network in such a distributed computing environment.

It will be appreciated by those skilled in the art that embodiments of the present description may be provided as a method, system, or computer program product. Thus, it will be apparent to those skilled in the art that the functional modules/units or controllers and associated method steps set forth in the above embodiments may be implemented in software, hardware, and a combination of software/hardware.

The acts of the methods, procedures, or steps described in accordance with the embodiments of the present invention do not have to be performed in a specific order and still achieve desirable results unless explicitly stated. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Various embodiments of the invention are described herein, but for brevity, description of each embodiment is not exhaustive and features or parts of the same or similar between each embodiment may be omitted. Herein, "one embodiment," "some embodiments," "example," "specific example," or "some examples" means that it is applicable to at least one embodiment or example, but not all embodiments, according to the present invention. The above terms are not necessarily meant to refer to the same embodiment or example. Those skilled in the art may combine and combine the features of the different embodiments or examples described in this specification and of the different embodiments or examples without contradiction.

The exemplary systems and methods of the present invention have been particularly shown and described with reference to the foregoing embodiments, which are merely examples of the best modes for carrying out the systems and methods. It will be appreciated by those skilled in the art that various changes may be made to the embodiments of the systems and methods described herein in practicing the systems and/or methods without departing from the spirit and scope of the invention as defined in the following claims.

Claims

1. A risk decision system with feature deriving function is characterized in that the wind decision system comprises a feature deriving platform, a model center subsystem, a business product management subsystem and a risk report generating subsystem, wherein,

The feature derivation platform is configured to extract basic variables according to the feed data, the three-party data and the internal data, generate derived variables according to the basic variables, and generate a data set according to the basic variables and the derived variables;

the risk report generation subsystem is configured to generate a risk report from the operational results of the business product.

2. The system of claim 1, wherein the business product management subsystem is further configured to perform a simulation test on the wind control model to generate a simulation test result.

3. The system of claim 1, wherein the incoming data comprises a transaction message, the feature derivation platform being specifically configured to automatically parse the transaction message to obtain basic variables in the transaction message; and generating derivative variables according to the configuration rules of the derivative variables based on the basic variables, wherein the transaction messages comprise personal second-generation credit message, enterprise credit message and third-party message.

4. The system of claim 1, wherein the business product management subsystem is configured to export the business product based on user selection and export a dataset and a wind control model associated with the business product.

5. The system of claim 1, wherein the business product management subsystem is configured to import an external business product according to user selection and import a dataset and a wind control model associated with the external business product.

6. The system of claim 1, wherein when the model center subsystem configures the wind control model associated with the business product, the model center subsystem is specifically configured to configure the wind control model with a policy management component comprising at least one of a decision set management component, a decision table management component, a decision tree management component, a scoring card management component, a machine learning model management component, and a decision flow management component.

7. The system of claim 1, wherein the risk report includes results of the transaction, feed data of the transaction, three-way data, internal data, and rules for hits.

8. The system of claim 1, further comprising a model optimization subsystem configured to optimize the wind control model according to an optimization mode, the optimization mode comprising at least one of an AB test mode, a champion challenger mode, a rule backtracking mode.

9. A risk decision method, characterized by being applied to a risk decision system with a feature derived function, wherein the risk decision system with the feature derived function comprises a feature derived platform, a model center subsystem, a business product management subsystem and a risk report generation subsystem, and the method comprises the following steps:

10. The method of claim 9, wherein the feature-derived risk decision system further comprises a model optimization subsystem, the method further comprising:

11. A computer readable storage medium having stored thereon a computer program, wherein the program when executed by a processor implements the method of any of claims 9-10.

12. An electronic device, comprising: a processor and a memory storing a computer program, the processor being configured to perform the method of any of claims 9-10 when the computer program is run.