CN115953233A - Risk assessment system - Google Patents
Risk assessment system Download PDFInfo
- Publication number
- CN115953233A CN115953233A CN202211642272.2A CN202211642272A CN115953233A CN 115953233 A CN115953233 A CN 115953233A CN 202211642272 A CN202211642272 A CN 202211642272A CN 115953233 A CN115953233 A CN 115953233A
- Authority
- CN
- China
- Prior art keywords
- node
- condition
- engine
- rule
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a risk assessment system, and relates to the field of wind control. The risk assessment system comprises a decision engine, a model engine and a feature engine; the decision engine is used for acquiring a calling request of the business system, reading a rule set according to the calling request, calling a model which is constructed in the model engine, calling the characteristics of the risk data to be evaluated extracted from the characteristic engine, generating a risk evaluation result according to the characteristics of the risk data to be evaluated, the model and the rule set, and sending the risk evaluation result to the business system; the model engine is used for constructing a model and sending the constructed model to the decision engine when receiving a model calling request of the decision engine; and the characteristic engine is used for extracting the characteristics of the risk data to be evaluated and sending the characteristics of the risk data to be evaluated to the decision engine when receiving a characteristic calling request of the decision engine. The risk identification method and the risk identification device are used for solving the problem of poor risk identification capability.
Description
Technical Field
The application relates to the field of wind control, in particular to a risk assessment system.
Background
With the professional development of black industries such as cheating and loan, counterfeiting personal data, fraud, money laundering and the like, a wind control means which is only configured by a rule strategy can be broken through repeated tests and heuristic characteristic value threshold values, and the hysteresis of rule adjustment causes the newly set strategy to have little effect and break through again in a short time. How to improve the risk recognition capability becomes a priority.
At present, bank finance type wind control products mainly make rule strategies through industry expert experience, and conduct risk control in a mode of calling client people to carry out credit and crediting auditors to audit client data.
Disclosure of Invention
The application provides a risk assessment system for solving the problem of poor risk identification capability.
The embodiment of the application provides a risk assessment system, which comprises a decision engine, a model engine and a feature engine;
the decision engine is used for acquiring a calling request of a business system, reading a rule set according to the calling request, calling a model which is constructed in the model engine, calling the characteristics of risk data to be evaluated extracted from the characteristic engine, generating a risk evaluation result according to the characteristics of the risk data to be evaluated, the model and the rule set, and sending the risk evaluation result to the business system;
the model engine is used for constructing a model and sending the constructed model to the decision engine when receiving a model calling request of the decision engine;
the feature engine is used for extracting features of risk data to be evaluated and sending the features of the risk data to be evaluated to the decision engine when receiving a feature calling request of the decision engine.
Optionally, the decision engine is specifically configured to obtain a pre-stored rule set from a rule decision repository according to the invocation request.
Optionally, the decision engine is further configured to obtain a rule set configured by a wind control expert through a visual background, and store the configured rule set in the rule decision repository.
Optionally, the decision engine is specifically configured to display a rule editing tool in the visualization background by using a canvas; and acquiring the operation of the wind control expert on the rule editing tool to generate the rule set.
Optionally, the decision engine is specifically configured to input the characteristics of the risk data to be evaluated into the model, execute the rule set, and generate the risk evaluation result.
Optionally, the decision engine is specifically configured to create a root node; reading a rule from the rule set; reading a condition from the rule; judging whether a type node corresponding to the parameter type in the condition exists or not, if not, adding a type node corresponding to the parameter type in the condition, and taking the type node corresponding to the parameter type in the condition as a child node of the root node;
the decision engine is specifically configured to determine whether a condition node corresponding to the condition exists, record a position of the condition node if the condition node corresponding to the condition exists, regard the condition as a condition node if the condition node corresponding to the condition does not exist, regard the condition node as a child node of a type node corresponding to a parameter type in the condition, and establish a memory table corresponding to the condition node according to the condition node;
the decision engine is specifically configured to read a next condition from the rule, and return to execute the step of determining whether a type node corresponding to a parameter type in the condition exists until each condition in the rule is executed;
the decision engine is specifically configured to use the first conditional node as a left input node of the first combination node, use the second conditional node as a right input node of the first combination node, and generate the first combination node; taking an (i-1) th combined node as a left input node of an ith combined node, and taking an (i + 1) th conditional node as a right input node of the ith combined node, wherein i is an integer greater than 1; taking the (n-2) th combined node as a left input node of the (n-1) th combined node, and taking the nth condition node as a right input node of the (n-1) th combined node, wherein the number of the condition nodes is n; the memory table of the left input node and the memory table of the right input node of each combined node are connected in series to form the memory table of the combined node;
the decision engine is specifically configured to read a next rule from the rule set, and return to the step of executing the step of reading a condition from the rule until the execution of each rule in the rule set is completed.
Optionally, the model engine is specifically configured to obtain sample data, extract features of the sample data, construct a model according to the features of the sample data, and determine that the constructed model can be used according to the test set and the time-span sample.
Optionally, the model engine is specifically configured to obtain offline data in a data warehouse, perform data preprocessing on the offline data in the data warehouse, and obtain the sample data, where the data preprocessing includes data extraction, cleaning, conversion, and loading.
Optionally, the feature engine is specifically configured to acquire the risk data to be evaluated sent by the business system, and obtain the feature of the risk data to be evaluated through calculation by a streaming calculation engine.
Optionally, the risk data to be evaluated includes business data and user behavior data.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages: in the embodiment of the application, the risk assessment system comprises a decision engine, a model engine and a feature engine, wherein the decision engine acquires a calling request of the business system, reads a rule set according to the calling request, calls a built model in the model engine, and calls features of risk data to be assessed extracted from the feature engine, generates a risk assessment result according to the features, the model and the rule set of the risk data to be assessed, and sends the risk assessment result to the business system. According to the risk assessment method and device, the decision engine is used for reading the rule set, the model engine is used for building the model, the feature engine is used for extracting features of the risk data to be assessed, the features, the model and the rule set of the risk data to be assessed are comprehensively considered, a risk assessment result is generated, even if a client can break through the wind control means of the rule set, the features of the risk data to be assessed and the wind control means of the model are difficult to break through, the risk identification capability can be improved, the possibility that the client cheats, such as loan and the like, are successfully implemented can be reduced, and the problem of poor risk identification capability is solved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a schematic structural diagram of a risk assessment system according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of a decision engine according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a decision engine according to an embodiment of the present application;
FIG. 4 is a schematic diagram of rules in an embodiment of the present application;
FIG. 5 is a schematic diagram of rule compilation and execution in one embodiment of the present application;
FIG. 6 is a schematic flow chart of the feature engine according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making creative efforts shall fall within the protection scope of the present application.
The application provides a risk assessment system, and applicable service scenes comprise cash credit service, mortgage credit service and supply chain financial service, and specifically comprise processes of incoming item primary screening, pre-credit final review, credit in-process monitoring, post-credit acceptance and the like.
In the embodiment of the present application, as shown in fig. 1, the risk assessment system includes a decision engine 101, a model engine 102, and a feature engine 103;
the decision engine 101 is configured to obtain a call request of the business system, read a rule set according to the call request, call a model that has been constructed in the model engine 102, and call a feature of risk data to be evaluated, which is extracted from the feature engine 103, generate a risk evaluation result according to the feature of the risk data to be evaluated, the model, and the rule set, and send the risk evaluation result to the business system;
the model engine 102 is used for constructing a model and sending the constructed model to the decision engine 101 when receiving a model calling request of the decision engine 101;
the feature engine 103 is configured to extract features of the risk data to be evaluated, and send the features of the risk data to be evaluated to the decision engine 101 when receiving a feature calling request of the decision engine 101.
According to the risk assessment method and device, the decision engine is used for reading the rule set, the model engine is used for building the model, the feature engine is used for extracting features of the risk data to be assessed, the features, the model and the rule set of the risk data to be assessed are comprehensively considered, a risk assessment result is generated, even if a client can break through the wind control means of the rule set, the features of the risk data to be assessed and the wind control means of the model are difficult to break through, the risk identification capability can be improved, the possibility that the client cheats, such as loan and the like, are successfully implemented can be reduced, and the problem of poor risk identification capability is solved.
In an embodiment, the decision engine is specifically configured to obtain a pre-stored rule set from the rule decision repository according to the invocation request.
The call request may include the input parameter type. For example, the call request includes a study calendar, a monthly salary, whether a car exists, whether a house exists, and whether the risk assessment result is to be qualified to apply for a credit card.
The decision engine obtains a pre-stored rule set from the rule decision repository according to the call request, or the decision engine obtains a rule set corresponding to a pre-stored parameter type from the rule decision repository according to the parameter type included in the call request.
In a specific embodiment, the decision engine is further configured to obtain a rule set configured by the wind control expert through the visual background, and store the configured rule set in the rule decision repository.
The wind control expert can configure the rule set through the visual background, and the method is convenient, quick and easy to operate.
In one embodiment, the decision engine is specifically configured to display the rule editing tool in a visualization backend by using a canvas (canvas); and acquiring the operation of the wind control expert on the rule editing tool to generate a rule set.
The traditional rule strategy is realized by means of code hard coding, and the realization mode has the following problems: hard coding realizes that the business rules are difficult to maintain; (2) hard coding implementation business rules are difficult to cope with changes; (3) When the business rule changes, the code needs to be modified, and the business rule can take effect after the service is restarted.
The rule strategy mode of the system is that a self-developed decision engine is used in the system, and the decision engine realizes the separation of the business decision from the application program code, receives data input, explains the business rule and makes the business decision according to the business rule. The decision engine is an input output platform.
After a decision engine is introduced into the system, the business rules are not resident in the system in the form of program codes any more, but are replaced by the decision engine for processing the rules, and the business rules are stored in a rule decision repository and are completely independent of programs. Business personnel can manage the business rules like managing data, such as querying, adding, updating, counting, submitting the business rules, and the like. The business rules are loaded into a decision engine for invocation by the business system.
The business rule change is inevitably caused in the operation process of the business system, the decision engine is provided, and the business rule part is realized by the decision engine, so that the business rule can be modified by the decision engine under the condition of normal operation of the system, and the business rule can be conveniently realized as required.
The decision engine brings the following benefits: (1) The business rules are separated from the system codes, and the centralized management of the business rules is realized; (2) The business rules can be expanded and maintained at any time under the condition of not restarting the service; (3) Business rules can be dynamically modified, so that the demand change can be quickly responded; (4) The decision engine is relatively independent and only concerns about the business rules, so that business analysts can also participate in editing and maintaining the business rules of the system; (5) reducing the cost and risk of hard-coding business rules; (6) And the rule editing tool provided by the decision engine is used, so that the complex business rule is simple to implement.
In a specific embodiment, the decision engine is specifically configured to input the features of the risk data to be evaluated into the model, execute the rule set, and generate a risk evaluation result.
In one embodiment, as shown in FIG. 2, a schematic workflow diagram of a decision engine is shown. In fig. 2, the wind control expert configures a rule set through a visual management background, and stores the rule set in a rule decision repository in a decision engine. And the user accesses the service system, the service system requests to call the decision engine, the decision engine calls the model engine, calls the feature engine, executes the rule set, generates an execution result, and sends the execution result to the service system to be displayed to the user.
In one embodiment, as shown in FIG. 3, a schematic workflow diagram of a decision engine is shown. In fig. 3, the service system inputs a call request to the decision engine, where the call request includes a study history, a monthly salary, whether there is a car, whether there is a house, and the like, and the decision engine executes the rule one, the rule two, the rule three, and the rule four stored in the rule decision repository through the computing engine to generate an execution result, where the execution result includes whether to be qualified to apply for a credit card, and outputs the execution result to the service system.
In one embodiment, the decision engine is specifically configured to create a root node; reading a rule from a rule set; reading a condition from the rule; judging whether a type node corresponding to the parameter type in the condition exists or not, if not, adding a type node corresponding to the parameter type in the condition, and taking the type node corresponding to the parameter type in the condition as a child node of the root node;
the decision engine is specifically used for judging whether a condition node corresponding to the condition exists or not, recording the position of the condition node if the condition node corresponding to the condition exists, taking the condition as the condition node if the condition node corresponding to the condition does not exist, taking the condition node as a child node of a type node corresponding to a parameter type in the condition, and establishing a memory table corresponding to the condition node according to the condition node;
the decision engine is specifically used for reading the next condition from the rule and returning to the step of judging whether the type node corresponding to the parameter type in the condition exists or not until all the conditions in the rule are executed;
the decision engine is specifically used for generating a first combined node by taking the first conditional node as a left input node of the first combined node and taking the second conditional node as a right input node of the first combined node; taking the (i-1) th combined node as a left input node of the ith combined node, and taking the (i + 1) th conditional node as a right input node of the ith combined node, wherein i is an integer greater than 1; taking the (n-2) th combined node as a left input node of the (n-1) th combined node, and taking the nth condition node as a right input node of the (n-1) th combined node, wherein the number of the condition nodes is n; the memory table of the left input node and the memory table of the right input node of each combination node are connected in series to form the memory table of the combination node;
and the decision engine is specifically used for reading the next rule from the rule set and returning to the step of reading a condition from the rule until the execution of each rule in the rule set is finished.
In one embodiment, as shown in FIG. 4, a schematic diagram of a rule is shown. Rule (P): is an inference statement composed of conditions and conclusions, generally expressed as if \ 8230and then. The if part of a rule is called LS and the then part is called RS. Condition (C): the "if statement" condition means the minimum atomic condition under which the division cannot be continued. In fig. 4, there is a rule P1, and the rule P1 includes LS and RS. LS includes four conditions C1, C2, C3, C4. C1 is (subject: this department), C2 is (monthly salary: 12000), C3 is (whether there is a house: yes), C4 is (whether there is a car: yes). The parameter in the condition C1 is the academic calendar, the parameter in the condition C2 is the monthly salary, the parameter in the condition C3 is whether a house exists, and the parameter in the condition C4 is whether a car exists. RS is action.
In one embodiment, as shown in FIG. 5, a schematic diagram of rule compilation and execution is shown. Fig. 5 includes a network a and a network B. The a network consists of type nodes (typenodes) and other conditional filter nodes (filternodes). The B network consists of two types of nodes, namely a bmemorynode and a join node, wherein the bmemorynode mainly stores a set after join, the join node comprises two input ports, two sets needing to be matched are respectively input, and the join node performs merging work and transmits the two sets to the next node.
In fig. 5, a root node (root node) is created, which is the entrance of the inference network. Taking rule P1, condition 1 (C1) is taken out of rule P1, where rule P1 includes at least one condition, and in FIG. 5, rule P1 includes four conditions C1, C2, C3, C4. The parameter type in condition 1 is checked and if it is a new type, a type node is added. In fig. 5, the parameter type in the condition C1 is the academic calendar, the parameter type in the condition C2 is the monthly salary, the parameter type in the condition C3 is whether there is a room, and the parameter type in the condition C4 is whether there is a car. Checking whether the node A corresponding to the condition 1 exists or not, and if so, recording the position of the node; if not, the condition 1 is used as an A node to be added into the network, and an A memory table is built according to the A node. Taking out the condition 2 from the rule P1, checking whether the node A corresponding to the condition 2 exists, and recording the position of the node if the node A exists; if not, the condition 1 is used as an A node to be added into the network, and an A memory table is built according to the A node. Taking out the condition 3 from the rule P1, checking whether the node A corresponding to the condition 3 exists, and recording the position of the node if the node A exists; and if not, adding the condition 3 as an A node into the network, and simultaneously establishing an A memory table according to the A node. Taking out the condition 4 from the rule P1, checking whether the node A corresponding to the condition 4 exists, and recording the position of the node if the node A exists; and if not, adding the condition 4 into the network as an A node, and simultaneously establishing an A memory table according to the A node. The four conditions in the rule P1 are executed.
In fig. 5, the left input node of the first combination node B (1) is a (1), and the right input node is a (2); the left input node of the second combination node B (2) is B (1), and the right input node is A (3); the left input node of the second combination node B (3) is B (2) and the right input node is a (4). Until all node bs have completed processing. And the memory table of the left input node and the memory table of the right input node of each combination node are connected in series to form the memory table of the combination node. And reading the next rule from the rule set, namely the rule P2, and executing the step of reading a condition from the rule P2 until all the rules in the rule set are executed.
In a specific embodiment, the model engine is specifically configured to obtain sample data, extract features of the sample data, construct a model according to the features of the sample data, and determine that the constructed model can be used according to a test set and an Out of Time sample (OOT).
The characteristic of the sample data is extracted, which can be to dig out the characteristic from the sample data, firstly perform data cleaning, then process and calculate the characteristic, and perform necessary filling on the missing value.
The cross-time sample is a verification sample of a time-testing window. Determining that the model that has been built can be used according to the test set and the cross-time samples means determining whether the model is available according to the performance evaluation of the model in the test set and the cross-time samples.
In a specific embodiment, the model engine is specifically configured to obtain offline data in the data warehouse, perform data preprocessing on the offline data in the data warehouse, and obtain sample data, where the data preprocessing includes data extraction, cleaning, conversion, and loading.
The data preprocessing may be an Extract-Transform-Load (ETL), which may be performed by extraction, washing, transformation, and loading.
The off-line data in the data warehouse is subjected to data preprocessing to obtain sample data, model training, model evaluation and model online are carried out in an off-line mode, and even if massive data does not exist, a regular model suitable for the moment can be provided.
In a specific embodiment, the feature engine is specifically configured to obtain risk data to be evaluated sent by the business system, and obtain features of the risk data to be evaluated through calculation by the streaming calculation engine.
Wherein, the stream type calculation engine refers to Flink. After the characteristics of the risk data to be evaluated are obtained through calculation of the stream type calculation engine, the characteristics of the risk data to be evaluated can be cached.
In one embodiment, the risk data to be assessed includes business data and user behavior data.
User behavior data may include, among other things, user identity, user phone, education, income, rosters, equipment, credit, social contact, etc.
In one embodiment, as shown in FIG. 6, a workflow diagram of a feature engine is shown. In fig. 6, the business system provides business data and user behavior data, the business data is stored in a relational database, e.g., mySQL or oracle, and the user behavior data is stored in a NoSQL database, e.g., HBase. Calculating according to the service data and the user behavior data through a stream type calculation engine to obtain a characteristic result, caching the characteristic result to obtain an aggregation index result, calling a decision engine by a service system, obtaining a dependence characteristic through a real-time characteristic engine by the decision engine, and reading the characteristic provided by the real-time characteristic engine from the aggregation index result.
In summary, in the embodiment of the present application, the risk assessment system includes a decision engine, a model engine, and a feature engine, where the decision engine obtains a call request of the business system, reads a rule set according to the call request, calls a model that has been built in the model engine, and calls features of risk data to be assessed extracted from the feature engine, generates a risk assessment result according to the features of the risk data to be assessed, the model, and the rule set, and sends the risk assessment result to the business system. According to the method and the device, the decision engine is used for reading the rule set, the model engine is used for building the model, the feature engine is used for extracting the features of the risk data to be evaluated, the features, the model and the rule set of the risk data to be evaluated are comprehensively considered, the risk evaluation result is generated, even if a client can break through the wind control means of the rule set, the features of the risk data to be evaluated and the wind control means of the model are difficult to break through, the risk identification capability can be improved, the possibility of successful implementation of fraudulent behaviors such as client cheating and loan can be reduced, and the problem of poor risk identification capability is solved.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A risk assessment system comprising a decision engine, a model engine, and a features engine;
the decision engine is used for acquiring a calling request of a business system, reading a rule set according to the calling request, calling a model which is constructed in the model engine, calling the characteristics of risk data to be evaluated extracted from the characteristic engine, generating a risk evaluation result according to the characteristics of the risk data to be evaluated, the model and the rule set, and sending the risk evaluation result to the business system;
the model engine is used for constructing a model and sending the constructed model to the decision engine when receiving a model calling request of the decision engine;
the feature engine is used for extracting features of the risk data to be evaluated and sending the features of the risk data to be evaluated to the decision engine when receiving a feature calling request of the decision engine.
2. The risk assessment system of claim 1, wherein the decision engine is specifically configured to obtain a pre-stored rule set from a rule decision repository according to the invocation request.
3. The risk assessment system of claim 2, wherein the decision engine is further configured to obtain a rule set configured by a wind control expert through a visual background, and store the configured rule set in the rule decision repository.
4. The risk assessment system of claim 3, wherein the decision engine, in particular for employing a canvas, displays a rule editing tool in the visualization backend; and acquiring the operation of the wind control expert on the rule editing tool to generate the rule set.
5. The risk assessment system of claim 4, wherein the decision engine is specifically configured to input features of the risk data to be assessed into the model, execute the rule set, and generate the risk assessment result.
6. The risk assessment system of claim 5, wherein the decision engine, in particular for creating a root node; reading a rule from the rule set; reading a condition from the rule; judging whether a type node corresponding to the parameter type in the condition exists or not, if not, adding a type node corresponding to the parameter type in the condition, and taking the type node corresponding to the parameter type in the condition as a child node of the root node;
the decision engine is specifically configured to determine whether a condition node corresponding to the condition exists, record a position of the condition node if the condition node corresponding to the condition exists, regard the condition as a condition node if the condition node corresponding to the condition does not exist, regard the condition node as a child node of a type node corresponding to a parameter type in the condition, and establish a memory table corresponding to the condition node according to the condition node;
the decision engine is specifically configured to read a next condition from the rule, and return to execute the step of determining whether a type node corresponding to a parameter type in the condition exists or not until each condition in the rule is executed;
the decision engine is specifically configured to use the first conditional node as a left input node of the first combination node, use the second conditional node as a right input node of the first combination node, and generate the first combination node; taking an (i-1) th combined node as a left input node of an ith combined node, and taking an (i + 1) th conditional node as a right input node of the ith combined node, wherein i is an integer greater than 1; taking the (n-2) th combined node as a left input node of the (n-1) th combined node, and taking the nth condition node as a right input node of the (n-1) th combined node, wherein the number of the condition nodes is n; the memory table of the left input node and the memory table of the right input node of each combination node are connected in series to form the memory table of the combination node;
the decision engine is specifically configured to read a next rule from the rule set, and return to the step of executing the step of reading a condition from the rule until the execution of each rule in the rule set is completed.
7. The risk assessment system of claim 1, wherein the model engine is specifically configured to obtain sample data, extract features of the sample data, construct a model according to the features of the sample data, and determine that the constructed model can be used according to the test set and the time-spanning sample.
8. The risk assessment system of claim 7, wherein the model engine is specifically configured to obtain offline data in a data warehouse, perform data preprocessing on the offline data in the data warehouse, and obtain the sample data, wherein the data preprocessing includes data extraction, cleaning, conversion, and loading.
9. The risk assessment system according to claim 1, wherein the feature engine is specifically configured to obtain the risk data to be assessed sent by the business system, and obtain the feature of the risk data to be assessed through calculation by a streaming computation engine.
10. The risk assessment system of claim 9, wherein the risk data to be assessed includes business data and user behavior data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211642272.2A CN115953233A (en) | 2022-12-20 | 2022-12-20 | Risk assessment system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211642272.2A CN115953233A (en) | 2022-12-20 | 2022-12-20 | Risk assessment system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115953233A true CN115953233A (en) | 2023-04-11 |
Family
ID=87289051
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211642272.2A Pending CN115953233A (en) | 2022-12-20 | 2022-12-20 | Risk assessment system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115953233A (en) |
-
2022
- 2022-12-20 CN CN202211642272.2A patent/CN115953233A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009174B (en) | Risk recognition model training method and device and server | |
| CN110188198B (en) | Anti-fraud method and device based on knowledge graph | |
| KR102061987B1 (en) | Risk Assessment Method and System | |
| Verbraken et al. | Development and application of consumer credit scoring models using profit-based classification measures | |
| US20200265511A1 (en) | Micro-Loan System | |
| CN108898476A (en) | A kind of loan customer credit-graded approach and device | |
| WO2020073727A1 (en) | Risk forecast method, device, computer apparatus, and storage medium | |
| CN110930038A (en) | Loan demand identification method, loan demand identification device, loan demand identification terminal and loan demand identification storage medium | |
| CN111861716A (en) | Method for generating monitoring early warning level in credit based on software system | |
| CN115545709A (en) | Abnormal fund allocation transaction identification method and device | |
| CN109271415B (en) | Data processing method and device for credit investigation database | |
| Dai et al. | Audit analytics: A field study of credit card after-sale service problem detection at a major bank | |
| CN117094764A (en) | Bank integral processing method and device | |
| CN110619564B (en) | Anti-fraud feature generation method and device | |
| CN115953233A (en) | Risk assessment system | |
| CN113177840A (en) | Client risk identification method and device | |
| CN114418767A (en) | Transaction intention identification method and device | |
| CN114626938A (en) | Intelligent decision engine, decision system and decision method | |
| CN114048330A (en) | Risk conduction probability knowledge graph generation method, device, equipment and storage medium | |
| Hayek et al. | Machine learning and external auditor perception: An analysis for UAE external auditors using technology acceptance model | |
| CN113256404A (en) | Data processing method and device | |
| CN111932368B (en) | Credit card issuing system and construction method and device thereof | |
| Mendes | Forecasting bitcoin prices: ARIMA vs LSTM | |
| CN114757723B (en) | Data analysis model construction system and method for resource element trading platform | |
| Rojos et al. | Process Mining: Research in Banking Operations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |