Risk control system and method based on multi-dimensional data model, and electronic equipment

Abstract

The invention provides a risk control system and method based on a multi-dimensional data model and electronic equipment, and the system comprises a data collection module which is used for collecting multi-dimensional user data from an internal data source and an external data source; the data preprocessing module is used for preprocessing the collected multi-dimensional user data and constructing a multi-dimensional data model based on the preprocessed data; wherein the multi-dimensional data model comprises user feature data divided according to different dimensions; the intelligent risk control model building module is used for building an intelligent risk control model based on the multi-dimensional data model; wherein the intelligent risk control model comprises at least one of a risk prediction model, a risk assessment model and a risk decision model; and the decision engine module is used for carrying out risk identification based on the intelligent risk control model and a preset rule to obtain a risk assessment result, and carrying out risk decision based on the risk assessment result. The data utilization rate and risk control efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent risk control, and in particular, to a risk control system, method, and electronic device based on a multi-dimensional data model. Background Art

[0002] With the increasing complexity of the financial market and the rapid growth of credit business, the traditional risk control system has been difficult to meet the risk management needs of financial institutions. Most of the existing risk control system products currently focus on data collection, data processing, and the development of rule engine technology. The system business logic is relatively single and the form is relatively traditional. The feedback cycle for complex custom risk disposal events is relatively long, and it seems powerless in the demand for rapid processing of multi-source data in Internet credit risk control technology. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a risk control system, method, and electronic device based on a multi-dimensional data model to improve data utilization and risk control efficiency.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0005] In the first aspect, the present invention provides a risk control system based on a multi-dimensional data model, including: a data acquisition module for collecting multi-dimensional user data from internal data sources and external data sources; a data preprocessing module for preprocessing the collected multi-dimensional user data and constructing a multi-dimensional data model based on the preprocessed data; wherein the multi-dimensional data model includes user feature data divided according to different dimensions; an intelligent risk control model construction module for building an intelligent risk control model based on the multi-dimensional data model; wherein the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model, and a risk decision model; a decision engine module for performing risk identification based on the intelligent risk control model and preset rules to obtain a risk assessment result, and performing risk decision based on the risk assessment result.

[0006] Optionally, the intelligent risk control model construction module includes: a model training unit for: determining risk feature variables based on the multi-dimensional data model to obtain a training sample set; training the training sample set using a machine learning algorithm to determine the weight and threshold of each risk feature variable to obtain an intelligent risk control model; evaluating and optimizing the intelligent risk control model to obtain a trained intelligent risk control model.

[0007] Optionally, the intelligent risk control model building module further includes: a model management unit, configured to manage the basic information of the intelligent risk control model, and display the grouping information, performance curve, and monitoring curve of the intelligent risk control model; wherein, the monitoring curve includes: the probability density curve of the risk characteristic variables of the intelligent risk control model, and the comparison curve of the probability density of the risk characteristic variables of the intelligent risk control model and the probability density of the training sample set.

[0008] Optionally, the intelligent risk control model processes the user characteristic data through built-in rules to obtain a risk assessment result; wherein, the built-in rules of the intelligent risk control model include at least one of the following: a rule set, a decision tree, a scoring card, and a decision flow engine.

[0009] Optionally, the decision engine module includes: a decision unit and a warning unit; the decision unit is configured to perform risk identification based on the intelligent risk control model and preset rules to obtain a risk assessment result, and perform a risk decision based on the risk assessment result; the warning unit is configured to give a warning based on the risk assessment result.

[0010] Optionally, it further includes: a monitoring and feedback module, configured to monitor the operating state of the risk control system and the risk assessment result, and display the operating state in a visual manner.

[0011] In a second aspect, the present invention provides a risk control method based on a multi-dimensional data model, which is applied to the risk control system based on a multi-dimensional data model according to any one of the first aspect. The method includes: collecting multi-dimensional user data from an internal data source and an external data source; preprocessing the collected multi-dimensional user data, and constructing a multi-dimensional data model based on the preprocessed data; wherein, the multi-dimensional data model includes user characteristic data divided according to different dimensions; building an intelligent risk control model based on the multi-dimensional data model; wherein, the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model, and a risk decision model; performing risk identification based on the intelligent risk control model and preset rules to obtain a risk assessment result, and performing a risk decision based on the risk assessment result.

[0012] Optionally, it further includes: monitoring the operating state of the risk control system and the risk assessment result, and displaying the operating state in a visual manner.

[0013] In a third aspect, the present invention provides an electronic device, including a processor and a memory, where the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the method according to any one of the second aspect.

[0014] Fourthly, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, it executes the steps of the method according to any one of the above-mentioned second aspects.

[0015] The present invention brings the following beneficial effects:

[0016] The risk control system, method and electronic device based on a multi-dimensional data model provided by the present invention include: a data acquisition module for acquiring multi-dimensional user data from internal and external data sources; a data preprocessing module for preprocessing the acquired multi-dimensional user data and constructing a multi-dimensional data model based on the preprocessed data. Among them, the multi-dimensional data model includes user feature data divided according to different dimensions; an intelligent risk control model construction module for building an intelligent risk control model based on the multi-dimensional data model. Among them, the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model and a risk decision model; a decision engine module for performing risk identification based on the intelligent risk control model and preset rules to obtain a risk assessment result, and performing risk decision based on the risk assessment result. Through the construction of the multi-dimensional data model, the above system deeply integrates and standardizes the multi-dimensional user data, can improve the breadth, depth and freshness of the data, thereby improving the utilization rate of the data, and providing more comprehensive and accurate data support for risk control; at the same time, through the construction of the intelligent risk control model, it can accurately identify and evaluate risks, and improve the accuracy and efficiency of risk control.

[0017] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0018] To make the above-mentioned objectives, features and advantages of the present invention more obvious and understandable, the following specific embodiments are given, and in conjunction with the accompanying drawings, the detailed description is as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a schematic structural diagram of a risk control system based on a multi-dimensional data model provided by an embodiment of the present invention;

[0021] Figure 2 A schematic structural diagram of another risk control system based on a multi-dimensional data model provided by an embodiment of the present invention;

[0022] Figure 3 A schematic diagram of a risk control system provided by an embodiment of the present invention;

[0023] Figure 4 A flowchart of a risk control method based on a multi-dimensional data model provided by an embodiment of the present invention;

[0024] Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0026] Currently, existing risk control platforms usually rely on a large amount of data for risk assessment and prediction, and the accuracy of the data directly determines the reliability of the model. If there are errors or omissions in the data, it will directly affect the accuracy and effectiveness of the risk control model.

[0027] Based on this, a risk control system, method, and electronic device based on a multi-dimensional data model provided by an embodiment of the present invention are used to improve data utilization and risk control efficiency.

[0028] For easy understanding of this embodiment, a risk control system based on a multi-dimensional data model disclosed in an embodiment of the present invention will be introduced in detail first. This system can be applied to various financial institutions, such as banks. This system uses mainstream Web front-end and back-end frameworks such as SpringBoot, SpringMVC, Mybatils, and REACT to build the system, realizes visual configuration of model files and decision flows based on human-computer interaction, and at the same time supports the import and direct application of training model files of types such as py and pmml; introduces the message queue Kafka component to push business data to the data mart in real time for real-time analysis of the data, and monitors the approval processes of various businesses through real-time indicator calculation, and completes the full life cycle monitoring and follow-up of business applications; adopts cluster deployment, which is convenient for horizontal expansion and system upgrade, and at the same time ensures business continuity and the needs of subsequent business incremental development.

[0029] See Figure 1A schematic structural diagram of a risk control system based on a multi-dimensional data model is shown, indicating that the method mainly includes the following parts:

[0030] The data acquisition module 101 is used to collect multi-dimensional user data from internal data sources and external data sources.

[0031] In one implementation, the data acquisition module 101 adopts a unified access and call input method, which is mainly divided into two parts: an input layer and data acquisition. The input layer includes a big data platform and an external data platform. The big data platform is the specific system carrier for all data within the bank, carrying the core data within the bank; the external data platform is the bridge connecting external data and business applications, responsible for integrating and managing external data from multiple sources and multiple structures.

[0032] During data acquisition, multi-dimensional user data such as credit information of the People's Bank of China, asset transactions, justice, and anti-fraud is obtained from data sources such as business systems and third-party interfaces. The data includes internal data and external data. At the same time, various types of data required in the risk control decision-making process are also widely collected to ensure the comprehensiveness and accuracy of the data.

[0033] Specifically, during data acquisition, the data acquisition module 101 can collect internal bank-owned data such as transactions and assets and liabilities, as well as external compliance data such as credit information of the People's Bank of China, tax, invoices, and third-party credit information. The three-party data acquisition component crawls various types of three-party data by providing a custom interface, thereby realizing flexible processing and standardized integration of different data sources. In addition, the data acquisition module 10 can not only process conventional structured data, but also parse and process unstructured data such as customer behavior data, device fingerprint data, and image and audio-visual data. In addition, according to decision-making needs, data can be compensated from other data sources. Both the service interface and the data compensation call interface comply with the relevant specifications of the central architecture control to ensure the compliance and security of the data.

[0034] In terms of data processing, the data acquisition module 10 can achieve dual processing of real-time data and batch data, clean and transform various types of data in existing databases, and push and load customer online behavior log information to the big data platform in real time to achieve data standardization and unification, and perform real-time online behavior analysis on real-time data. The batch data is summarized to the big data platform at a fixed time every day and stored in the data model library, also achieving data standardization and unification, and further processing the data.

[0035] In the embodiment of the present invention, for data acquisition, the data acquisition module 10 has the ability to fuse multi-source data, adopts a modular management process, reduces the data access time and cost by integrating internal and external data of the business party, breaks through the data islands of the business party, provides risk data with unified, comprehensive, and comprehensive caliber, and improves the data management efficiency.

[0036] The data preprocessing module 102 is used to preprocess the collected multi-dimensional user data and construct a multi-dimensional data model based on the preprocessed data.

[0037] In one implementation, the above data collection module 101 collects multi-dimensional user data such as personal credit investigation, asset transactions, judiciary, anti-fraud, etc. from internal and external data sources. The data preprocessing module 102 cleans and stores the data. The data preprocessing module 102 can directly return the data to the decision flow or push it to the cache database for cache calculation.

[0038] Specifically, data cleaning refers to eliminating or filling in business data that does not meet actual needs according to certain rules, mainly including the following steps:

[0039] (1) View the data to determine the structure, features, attributes, etc. of the data, and prepare for subsequent data cleaning work.

[0040] (2) Process missing values: For the missing values in the data, they can be processed by deletion, filling in the average value, mode or specific value, etc. The selection of specific methods can be determined according to the actual situation of the data and the requirements of subsequent analysis.

[0041] (3) Process outliers: For outliers, they can be identified by methods such as box plots and the 3σ principle, and corresponding processing measures can be taken, such as deletion, replacement or correction, etc.

[0042] (4) Data verification: Verify the data to ensure the accuracy, integrity and consistency of the data, including: checking whether the data format is correct, whether the field values are legal, and whether there are logical contradictions between the data, etc.

[0043] In addition, the data preprocessing module 102 can generate risk feature variables by calling operators with different parameters; it can also further derive the basic risk feature variables through mathematical operations and logical operations, greatly improving the standardization and efficiency of variable processing. Specifically, the data mining and derivation include: (1) Pre-configure functions and SOL logic for data derivation, and ensure the accuracy of the configured logic through methods such as SQL comparison, trial calculation and trial operation. (2) Save and configure risk feature variables through the risk feature library. (3) Perform various data mining processes on the collected data, such as association analysis, classification, clustering, regression prediction, etc.

[0044] After completing the preprocessing of the multi-dimensional user data, a multi-dimensional data model can be constructed based on the preprocessed data, including user feature data in multiple dimensions such as time, geographical location, product, customer, etc., to support complex data analysis and queries.

[0045] The intelligent risk control model construction module 103 is used to build an intelligent risk control model based on a multi-dimensional data model; among them, the intelligent risk control model includes one or more of the following: a risk prediction model, a risk assessment model, and a risk decision model.

[0046] In one implementation, the intelligent risk control model construction module 103 can use intelligent algorithms such as machine learning and deep learning to build models such as a risk prediction model, a risk assessment model, and a risk decision model, so as to achieve accurate identification and assessment of risks.

[0047] In specific implementation, machine learning algorithms such as logistic regression, support vector machine, adaptive boosting, and decision tree are used to build models such as a user's repayment ability model, a credit investigation model, an identity authentication model, a fraud model, a risk pricing model, a quota assessment model, and a credit scoring model, forming a multi-dimensional model library. Combining expert experience, an intelligent risk control model that meets financial institutions is constructed, and centralized management of the intelligent risk control model is carried out, including: processes such as model construction, verification, deployment, and optimization.

[0048] The decision engine module 104 is used to perform risk identification based on the intelligent risk control model to obtain a risk assessment result, and perform risk decision-making based on the risk assessment result.

[0049] In one implementation, the decision engine module 104 is used for the whole-process risk assessment and decision-making of credit business, applies the whole-life cycle risk control scenarios of online data loan product credit risk and fraud risk, embeds the intelligent risk control model service into the online and offline processes of credit business, and realizes intelligent access and intelligent approval, improving the approval efficiency. Specifically, the decision engine module 104 includes a rule engine, a monitoring center, data access, and corresponding system management, etc. It is mainly responsible for the configurability of business rules and risk rules, performs model management and strategy allocation of risk rules, formulates monitoring rules and monitoring thresholds for pre-transaction, and at the same time provides real-time risk decision-making for in-process risk rules, supporting more complex and flexible settings, greatly improving the flexibility, intelligence, and risk prevention and control ability of the system.

[0050] In specific implementation, the intelligent risk control model processes user feature data through built-in rules to obtain a risk assessment result; among them, the built-in rules of the intelligent risk control model include at least one of the following: rule set, decision tree, scoring card, decision flow engine. The rule engine presets a variety of rules for decision-making, specifically performs different branch combinations and associations on the business rules abstracted from complex business logics, and then performs progressive operations on the layers of rules to obtain a product that outputs a decision result. Among them, the rule engine includes rule function components such as data model, function variable, decision tree, decision matrix, decision table, scoring card, rule, rule function, rule set, decision flow, etc., and at the same time is equipped with corresponding functions such as single-item test, batch test, online deployment, etc.

[0051] The above-mentioned risk control system based on a multi-dimensional data model provided by the present invention, through the construction of the multi-dimensional data model, deeply integrates and standardizes multi-dimensional user data, can improve the breadth, depth and freshness of the data, thereby improving the utilization rate of the data, and providing more comprehensive and accurate data support for risk control; at the same time, through the construction of the intelligent risk control model, it can accurately identify and evaluate risks, and improve the accuracy and efficiency of risk control.

[0052] In one implementation manner, the intelligent risk control model building module 103 includes: a model training unit, which is used to: determine risk feature variables based on the multi-dimensional data model to obtain a training sample set; train the training sample set using a machine learning algorithm to determine the weight and threshold of each risk feature variable, and obtain an intelligent risk control model; evaluate and optimize the intelligent risk control model to obtain a trained intelligent risk control model.

[0053] In specific implementation, taking the credit scoring model as an example, the training process of the model training unit is introduced in detail. The construction of the model mainly includes:

[0054] (1) Model selection: Select a suitable model for building the credit scoring model according to the characteristics of the data and business requirements. Commonly used models include logistic regression, decision tree, neural network, etc.

[0055] (2) Model training: Use the selected model to train the data to determine the weight and threshold of each risk feature variable. During the training process, the characteristics and distribution of the data should be fully considered to avoid overfitting or underfitting problems.

[0056] After the model is constructed, it is necessary to evaluate and verify the model. The steps are as follows:

[0057] (1) Model evaluation: Evaluate the trained model, including aspects such as accuracy and stability. Commonly used evaluation indicators include ROC curve, KS index, discrimination, goodness-of-fit curve, confusion matrix, etc. Through evaluation, the performance of the model can be understood for subsequent optimization.

[0058] (2) Model optimization: Optimize the model according to the evaluation results, including adjusting parameters, adding or reducing features, etc. Common parameter tuning methods include grid search, random search, Bayesian optimization, gradient descent method, and ensemble learning, etc. These methods are used to determine the best parameter combination of the model, enabling the model to better adapt to the risk control scenario, so as to improve the accuracy and stability of the model. Optimization should be comprehensively considered in combination with the business background and data characteristics to improve the performance and practicality of the model.

[0059] After completing the configuration of the intelligent risk control model and determining the thresholds of each intelligent risk control model, offline data and online data can be used to conduct trial calculations on the model. Offline trial calculation supports calculating the results of using the new model for the existing business, while online trial calculation is to send the current business concurrently to the new model for measurement. Based on the trial calculation report, business personnel can conduct a full-scale evaluation of the new intelligent risk control model from risk data to the final model results.

[0060] Generally, most offline calculation tasks are scheduled at T+1 or T+N. Due to the limitation of computing power, only a very small number adopt hourly task scheduling. In the embodiments of the present invention, by integrating a streaming computing engine, the data processing link can be broadened, supporting hourly and minute-level task scheduling calculation jobs, greatly improving the construction of data dimensions for business analysis. At the same time, it supports accessing transaction data from various trading channels in an asynchronous or synchronous manner, detecting through the access strategy engine of the data service specification of the anti-fraud platform, and realizing the access and system-level integration of channel transaction data.

[0061] In the embodiments of the present invention, users can develop intelligent risk control models online through the intelligent risk control model building module 102 without relying on the traditional offline development environment. Visualize the deployment of models and strategies such as pre-loan application scoring, customer segmentation groups, application anti-fraud, behavior scoring, post-loan quota adjustment, risk warning, and first collection, and complete real-time decision-making through the decision flow engine, rule engine, and model engine. Online development can speed up the process from model design to deployment and improve development efficiency. At the same time, users can intuitively build, debug, and optimize the model through the tools and interfaces provided by the platform, reducing the operation difficulty.

[0062] Furthermore, the intelligent risk control model building module 102 further includes: a model management unit, which is used to manage the basic information of the intelligent risk control model, and display the grouping information, performance curve, and monitoring curve of the intelligent risk control model; wherein, the monitoring curve includes: the probability density curve of the risk characteristic variables of the intelligent risk control model, and the comparison curve of the probability density of the risk characteristic variables of the intelligent risk control model and the probability density of the training sample set.

[0063] In specific implementation, after the model is evaluated and verified, the model needs to be managed, mainly including: the serial number of the model, name, model type, current status (whether it is in effect), model access conditions, the weight of the model in the review, the series of actions related to the model, the algorithm used by the model, the author of the model, the training objective of the model, the model discrimination index, the model training date, and the release date, etc. In addition, the list of risk characteristic variables on which the model depends can also be displayed in the model management unit, specifically including the following content:

[0064] (1) The grouping information of the model, and what is the weight of the model in its current grouping;

[0065] (2) The performance curves of the model, including the KS curve, AUC / ROC curve, PR curve, and Lift curve;

[0066] (3) The model monitoring curves, including the probability density curve of the risk characteristic variables on which the model depends, and the comparison between the probability density of each current risk characteristic variable and the probability density curve of the training sample set for model training.

[0067] Furthermore, the model management unit embeds multiple model building patterns, and the backend provides a powerful model training resource configuration model, which can support the training and deployment of large language models, and at the same time support functions such as multi-person collaborative modeling, distributed debugging, and review and verification. The model management unit can perform full life cycle management on rules and models, including various stages such as creation, editing, testing, deployment, monitoring, and retirement, so as to ensure the version consistency of the model and reduce the maintenance cost. At the same time, it can also record the historical change situations of each model, which is convenient for tracking and auditing.

[0068] In one implementation, the decision engine module 104 includes: a decision-making unit and a warning unit; the decision-making unit is used to perform risk identification based on the intelligent risk control model and preset rules to obtain a risk assessment result, and perform risk decision-making based on the risk assessment result; the warning unit is used to give a warning based on the risk assessment result.

[0069] In specific implementation, a convenient and easy-to-use visual editing interface can be provided for users, and various combinations of judgment logics and process links can be configured to realize the flexible configuration and combination of preset rules (rule sets, scorecards, decision tables, etc.), customize decision-making behaviors, and take the decision flow as the main line for different business scenarios, and connect basic decision-making components such as scorecards, decision trees, and decision tables in series to realize a complete risk control process. Suppose a bank needs to evaluate the credit risk of an applicant, then the constructed credit risk prediction model can be used to automatically evaluate the credit risk of the applicant according to preset rule sets, decision trees, scorecards and other rules.

[0070] The above-mentioned decision engine module 104 provided by the embodiments of the present invention further includes a decision management service, which can dynamically deploy rules, policies, and models, that is, update or replace existing rules, policies, and models without restarting the system or interrupting the service. Dynamic deployment can reduce system downtime and improve the availability and stability of the system. At the same time, the deployment process can be monitored in real time, and the deployment progress and results can be displayed in real time, facilitating users to discover and solve problems in a timely manner.

[0071] The above-mentioned decision management service can also perform value assignment adjustment on the parameters in the rules, policies, and models, that is, users can optimize the effects of the rules, policies, and models by modifying the parameter values. Value assignment adjustment enables users to more flexibly control the behavior of the rules, policies, and models to meet different business needs. At the same time, it can also provide suggestions and guidelines for parameter adjustment to help users set parameter values more reasonably.

[0072] Furthermore, the above-mentioned decision management service further includes reusable components such as metrics, rules, and decision flows. Users can build complex rules, policies, and models by editing these components, which can be used to back up historical model versions, establish a general model library, and implement the full life cycle process management of risk models; the components support drag-and-drop editing, reducing the learning cost and usage difficulty of users. At the same time, the deployment and operation of these components can also be monitored to ensure that they can execute correctly and produce the expected effects. Business personnel can quickly configure rule combinations according to changes in internal and external risk characteristics to achieve "plug-and-play", greatly shortening the development cycle of monitoring rules.

[0073] With the support of the above functions, the risk control system can provide users with efficient decision management services, provide standardized decision result returns according to different business scenarios and decision operation results, and provide decision-making basis for interruptions and branches in the business process. Users can quickly build and optimize risk control models through the platform to improve decision-making efficiency and accuracy. At the same time, it can also display real-time risk decision results and early warning information to help users make correct decisions and response measures in a timely manner.

[0074] In addition, the decision management service also supports the following function operations:

[0075] (1) Adding real-time derived risk variables to the decision flow: After business personnel derive new risk characteristic variables, the risk characteristic variables can be immediately selected in the decision flow without waiting for the synchronization process. Using a stream processing engine, by subscribing to calculation scripts, when data objects are pushed to the stream processing engine, the specified calculation scripts are executed, and the calculated metrics can be placed in the cache to improve the efficiency of real-time calculation.

[0076] (2) Decision flow threshold templatized configuration: After the risk decision model configuration is finalized, the system will automatically generate an encrypted configuration document. Business personnel can fill in the configuration document offline and upload it to take effect, which is convenient for the later adjustment and maintenance of the model.

[0077] (3) Adopt real-time streaming processing technology and high-performance distributed microservice architecture. Conduct distributed cache management on behavioral feature indicators, real-time transaction flows, rule factors, etc., greatly reducing the dependence on database reads. To meet the system performance pressure brought by high-concurrency business scenarios and the rapid growth of the business, it can quickly respond to the calculation tasks of thousands of decision models / rules, ensure the efficient operation of core processing modules such as its own statistical engine and rule engine, and better support high-concurrency processing capabilities.

[0078] Through the online development ability of the model, unified management of the entire life cycle of rules and models, dynamic deployment monitoring, assignment-based adjustment, and reusable components, the embodiments of the present invention provide users with efficient, flexible, and reliable decision management services, enabling the system to better meet the needs of users in decision management and improving business efficiency and competitiveness.

[0079] In one implementation, as shown in Figure 2 the above system further includes: a monitoring and feedback module 105, which is used to monitor the operating status and risk assessment results of the risk control system and display the operating status in a visual manner.

[0080] In specific implementation, the monitoring and feedback module 105 can conduct full-process risk monitoring on credit business, including real-time monitoring in the pre-loan, in-loan, and post-loan stages, issue early warnings in a timely manner according to the risk assessment results, and take differentiated risk prevention and control measures, such as increasing credit limits, reducing interest rates, restricting credit limits, increasing interest rates, etc., so as to achieve early detection, early warning, and early disposal of risks.

[0081] Specifically, the operating status of risk control attributes, variables, rules, and models can be monitored and quickly warned through monitoring reports. Through a graphical statistical analysis interface, the success rate monitoring statistics of each decision and multi-dimensional intelligent statistical analysis of models and touched decisions can be realized, etc., so as to be able to evaluate the effectiveness of model strategies in a timely manner, discover system problems, fraud behaviors, credit risks, etc. in a timely manner, and reduce operational risks. The monitoring reports specifically include the following:

[0082] (1) The monitoring report can conduct abnormal warnings, and the warning rules can be configured through SQL. At the same time, it can conduct flexible report custom settings in the interface and flexibly display fixed reports;

[0083] (2) Monitoring reports can focus on the entire process and life cycle of risk control strategy system construction. From strategy development, iterative optimization, monitoring evaluation and offline, risk control personnel can quickly and comprehensively understand the overall picture of business risk control.

[0084] The monitoring and feedback module 105 also builds a full-link log, which can realize the implementation and connection of logs of all execution links of a single risk control model, and continuously monitor the running status of the model, and conduct real-time monitoring of core indicators such as the number of calls, response time, and distribution of variables.

[0085] The monitoring and feedback module 105 can also perform flow monitoring, efficiency monitoring, data monitoring, decision result monitoring, anomaly detection and model effect monitoring. Among them, flow monitoring and efficiency monitoring act on the system itself and are the guarantee of stable system operation; data monitoring and decision result monitoring focus on business, monitor the process indicators and results of business processing, and can timely and effectively determine business anomalies to ensure business stability; model monitoring and anomaly detection focus on real-time analysis. The former can observe the model effect in time to ensure that the model is running as expected, while the latter is biased towards anomaly interception. If a user is associated with other users or discovers anomalies after making decisions, the anomaly detection module will promptly notify the subsequent business process, intercept the user in time, and recover possible losses.

[0086] The monitoring and feedback module 105 can also provide a risk dashboard, showing the risk region and the amount of risk loss in the form of a map, to help business personnel analyze the overall risk for key prevention and control and risk adjustment. At the same time, a business dashboard is provided to display information such as risk type, customer classification, customer rating and routing results in the form of a column, pie chart, bar chart, line chart, etc., to help business personnel analyze the overall business for key prevention and control and business adjustment.

[0087] See also Figure 3 A schematic diagram of a risk control system is shown, which can be applied to scenarios such as credit card products, online loan products, and channel backflow products. It collects external data (user information, behavior information, transaction information, etc.) and internal data (PBOC credit, provident fund, industry and commerce, courts, etc.) to build a risk feature library, and performs low-risk identification and decision-making by applying intelligent risk control models such as anti-fraud models, pre-loan risk control models, mid-loan pricing and risk control models, and post-loan early warning management models, thereby achieving closed-loop risk management of pre-loan approval, mid-loan monitoring, and post-loan collection.

[0088] The risk control system provided by the embodiments of the present invention integrates a large amount of non-financial and financial data online and offline for multi-dimensional data modeling; combines third-party data such as credit investigation and industry and commerce to build a scientific and reasonable intelligent risk control model, and identifies fraud risks and credit risks through the intelligent risk model; uses various strategy models (black and white list strategies, identity verification strategies, approval strategies, scoring strategies, etc.) to achieve a risk closed-loop management for pre-loan approval, in-loan monitoring, and post-loan collection; through the microservice cluster technology architecture, it realizes the collection and real-time processing of big data, improves the data integration ability; constructs risk control decisions, realizes the flexible configuration and online of business rules, meets the requirements of different scenarios such as risk control, anti-fraud, and marketing, and can help financial institutions give play to the value of data assets and accelerate digital transformation. It comprehensively improves the customer experience, and then improves the operation level and core competitiveness of financial institutions.

[0089] For the risk control system based on the multi-dimensional data model provided in the foregoing embodiments, the embodiments of the present invention also provide a risk control method based on the multi-dimensional data model. Refer to Figure 4 The flowchart of a risk control method based on a multi-dimensional data model shown in the figure shows that the method mainly includes the following steps S401 to step S404:

[0090] Step S401: Collect multi-dimensional user data from internal data sources and external data sources;

[0091] Step S402: Preprocess the collected multi-dimensional user data, and build a multi-dimensional data model based on the preprocessed data; wherein, the multi-dimensional data model includes user feature data divided according to different dimensions;

[0092] Step S403: Build an intelligent risk control model based on the multi-dimensional data model; wherein, the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model, and a risk decision model;

[0093] Step S404: Perform risk identification based on the intelligent risk control model to obtain a risk assessment result, and make a risk decision based on the risk assessment result.

[0094] The above-mentioned risk control method based on the multi-dimensional data model provided by the present invention, through the construction of the multi-dimensional data model, deeply integrates and standardizes the multi-dimensional user data, can improve the breadth, depth and freshness of the data, thereby improving the utilization rate of the data, and providing more comprehensive and accurate data support for risk control; at the same time, through the construction of the intelligent risk control model, it can accurately identify and evaluate risks, and improve the accuracy and efficiency of risk control.

[0095] In one implementation manner, the above method further includes: monitoring the operating state of the risk control system and the risk assessment result, and displaying the operating state in a visual manner.

[0096] It should be noted that for the method provided in the embodiments of the present invention, its implementation principle and the resulting technical effects are the same as those of the foregoing system embodiments. For the sake of brief description, for the parts not mentioned in the method embodiments, reference may be made to the corresponding content in the foregoing system embodiments.

[0097] The embodiments of the present invention further provide an electronic device. Specifically, the electronic device includes a processor and a storage device; a computer program is stored on the storage device, and when the computer program is run by the processor, it executes the method described in any one of the above embodiments.

[0098] Figure 5 FIG. 9 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device 100 includes: a processor 50, a memory 51, a bus 52, and a communication interface 53. The processor 50, the communication interface 53, and the memory 51 are connected through the bus 52; the processor 50 is used to execute an executable module stored in the memory 51, such as a computer program.

[0099] Among them, the memory 51 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 53 (which may be wired or wireless), a communication connection is established between this system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0100] The bus 52 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 5 only a bidirectional arrow is used in FIG. 9, but it does not mean that there is only one bus or one type of bus.

[0101] Among them, the memory 51 is used to store a program. After receiving an execution instruction, the processor 50 executes the program. The method executed by the device defined by the flow process disclosed in any one of the foregoing embodiments of the present invention can be applied to the processor 50 or implemented by the processor 50.

[0102] The processor 50 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 50 or the instructions in the form of software. The above-mentioned processor 50 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 51, and the processor 50 reads the information in the memory 51 and combines its hardware to complete the steps of the above method.

[0103] The computer program product of the readable storage medium provided by the embodiments of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For the specific implementation, reference can be made to the foregoing method embodiments, which will not be elaborated here.

[0104] If the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0105] Finally, it should be noted that the above-mentioned embodiments are only specific embodiments of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions recorded in the foregoing embodiments, or can easily conceive of changes, or make equivalent replacements for some of the technical features; and these modifications, changes, or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A risk control system based on a multidimensional data model, characterized in that: include: A data collection module is used to collect multi-dimensional user data from internal and external data sources; A data preprocessing module, used to preprocess the collected multi-dimensional user data, and construct a multi-dimensional data model based on the preprocessed multi-dimensional user data; wherein the multi-dimensional data model includes user feature data divided according to different dimensions; An intelligent risk control model building module, used to build an intelligent risk control model based on the multidimensional data model; wherein the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model and a risk decision model; The decision engine module is used to identify risks based on the intelligent risk control model to obtain risk assessment results, and make risk decisions based on the risk assessment results.

2. The system according to claim 1, characterized in that The intelligent risk control model building module includes: a model training unit, which is used to: Determine risk characteristic variables based on the multidimensional data model to obtain a training sample set; Using a machine learning algorithm to train the training sample set, determine the weight and threshold of each risk feature variable, and obtain an intelligent risk control model; The intelligent risk control model is evaluated and optimized to obtain a trained intelligent risk control model.

3. The system according to claim 2, characterized in that The intelligent risk control model building module also includes: a model management unit, which is used to manage the basic information of the intelligent risk control model, and to display the grouping information, performance curve and monitoring curve of the intelligent risk control model; wherein the monitoring curve includes: a probability density curve of the risk characteristic variables of the intelligent risk control model, and a comparison curve of the probability density of the risk characteristic variables of the intelligent risk control model and the probability density of the training sample set.

4. The system according to claim 1, characterized in that The intelligent risk control model processes the user feature data through built-in rules to obtain a risk assessment result; wherein the built-in rules of the intelligent risk control model include at least one of the following: a rule set, a decision tree, a scoring card, and a decision flow engine.

5. The system according to claim 1, characterized in that The decision engine module includes: a decision unit and an early warning unit; The decision-making unit is used to identify risks based on the intelligent risk control model and preset rules to obtain risk assessment results, and make risk decisions based on the risk assessment results; The early warning unit is used to issue an early warning based on the risk assessment result.

6. The system according to claim 1, characterized in that Also includes: The monitoring and feedback module is used to monitor the operating status of the risk control system and the risk assessment results, and to display the operating status in a visual manner.

7. A risk control method based on a multidimensional data model, characterized in that: Applied to the risk control system based on the multidimensional data model according to any one of claims 1 to 6, the method comprising: Collect multi-dimensional user data from internal and external data sources; Preprocessing the collected multi-dimensional user data, and constructing a multi-dimensional data model based on the pre-processed data; wherein the multi-dimensional data model includes user feature data divided according to different dimensions; Building an intelligent risk control model based on the multidimensional data model; wherein the intelligent risk control model includes at least one of the following: a risk prediction model, a risk assessment model, and a risk decision model; Risk identification is performed based on the intelligent risk control model to obtain a risk assessment result, and risk decisions are made based on the risk assessment result.

8. The method according to claim 7, characterized in that Also includes: The operating status of the risk control system and the risk assessment results are monitored, and the operating status is displayed in a visual manner.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the method according to any one of claims 7 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 7 to 8 are performed.

Images