Decision model generation method and device based on data model and strategy model

Abstract

The invention provides a decision model generation method and device based on a data model and a strategy model, and the method comprises the steps: obtaining a strategy model and a data model through an intelligent decision engine, inputting the data model into the strategy model, and transmitting a decision model generation instruction to a strategy execution module; the decision model generation instruction is used for instructing the main strategy execution module and the sub-strategy execution module to screen out a target data set from the data model according to the data demand label; and establishing a complete decision model according to the main strategy model and the first target data set corresponding to the main strategy model, and the sub-strategy model and the second target data set corresponding to the sub-strategy model. Therefore, in the application, the intelligent decision engine decouples the data model and the strategy model to realize sharing of the data model between the main strategy model and the sub-strategy model; and sub-strategy input / output parameters are automatically derived, so that the workload of manual configuration is reduced, and the error rate is reduced.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and apparatus for generating decision models based on data models and strategy models. Background Technology

[0002] Existing intelligent decision-making systems primarily rely on rule engines or machine learning models. Rule engines are a fundamental service in enterprise applications; they use predefined semantic modules to write business decisions and provide feedback on decision requests based on pre-defined inference rules. Financial institutions typically use centrally deployed enterprise-level rule engines to uniformly process business decision requests for financial transactions.

[0003] In traditional decision-making methods, sub-strategies require manual declaration of input and output parameters, which is prone to omissions and errors. Sub-strategies and parent strategies cannot achieve synchronized parameter settings, resulting in sub-strategies redefining parameters already existing in the parent strategy, or modifications to parameters in the parent or sub-strategies cannot be automatically synchronized, leading to the inefficiency of "one modification, multiple adaptations" and inconsistencies in strategy logic due to untimely synchronization. Summary of the Invention

[0004] This application provides a method and apparatus for generating a decision model based on a data model and a strategy model. The intelligent decision engine reduces maintenance costs by decoupling the data model and the strategy model and enabling bidirectional sharing of the data model between the main strategy model and the sub-strategy model; it also automatically derives the input / output parameters of the sub-strategy, reducing the workload of manual configuration and lowering the error rate.

[0005] Firstly, this application provides a decision model generation method based on a data model and a strategy model, applied to an intelligent decision engine in an intelligent decision system. The intelligent decision system includes the intelligent decision engine and a strategy execution module. The strategy execution module includes a main strategy execution module and a sub-strategy execution module. The method includes: acquiring a strategy model, which includes a main strategy model and a sub-strategy model, each strategy model containing data requirement labels; acquiring a data model, which is associated with the main strategy model, the data model including multiple datasets, each dataset having a uniquely identified data label, the data label having a mapping relationship with the data requirement labels; inputting the data model into the strategy model and sending a decision model generation instruction to the strategy execution module, the decision model generation instruction instructing the main strategy execution module and the sub-strategy execution module to filter target datasets from the data model according to the data requirement labels; and establishing a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

[0006] Secondly, this application provides a decision-making method based on a data model and a strategy model, applied to an intelligent decision engine in an intelligent decision-making system. The intelligent decision-making system includes the intelligent decision engine and a strategy execution module. The strategy execution module includes a main strategy execution module and a sub-strategy execution module. The method includes: receiving a decision request from a terminal device; obtaining a strategy model according to the decision request, the strategy model including a main strategy model and a sub-strategy model, each strategy model containing a data requirement label; obtaining a data model according to the main strategy model, the data model being associated with the main strategy model, the data model including multiple datasets, each dataset having a uniquely identified data label, the data label having a mapping relationship with the data requirement label; inputting the data model into the strategy model and sending a decision model generation instruction to the strategy execution module, the decision model generation instruction instructing the main strategy execution module and the sub-strategy execution module to filter target datasets from the data model according to the data requirement labels; establishing a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset; receiving strategy data from the terminal device; inputting the strategy data into the decision model to obtain a decision result.

[0007] Thirdly, this application provides a device for generating a decision model based on a data model and a strategy model, applied to an intelligent decision engine. The device includes: an acquisition unit for acquiring a strategy model, the strategy model including a main strategy model and a sub-strategy model, each strategy model containing data requirement labels; and acquiring a data model associated with the main strategy model, the data model including multiple datasets, each dataset having a uniquely identified data label, the data label having a mapping relationship with the data requirement labels; a processing unit for inputting the data model into the strategy model and sending a decision model generation instruction to the strategy execution module, the decision model generation instruction instructing the main strategy execution module and the sub-strategy execution module to filter target datasets from the data model according to the data requirement labels; and establishing a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

[0008] Fourthly, this application provides an intelligent decision engine, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the step instructions of the method as described in any one of the first aspects when it invokes the computer program in the memory.

[0009] Fifthly, this application provides an intelligent decision-making system, including an intelligent decision-making engine, a terminal device, and a strategy execution module. The strategy execution module includes a main strategy execution module and a sub-strategy execution module. The intelligent decision-making engine is used to execute the step instructions in the method as described in any one of the first aspects.

[0010] As can be seen, in this embodiment, the intelligent decision engine acquires a strategy model and a data model. The strategy model includes a main strategy model and sub-strategy models. Each strategy model contains data requirement labels. The data model is associated with the main strategy model and includes multiple datasets. Each dataset has a uniquely identified data label, and there is a mapping relationship between the data labels and the data requirement labels. The data model is then input into the strategy model, and a decision model generation instruction is sent to the strategy execution module. The decision model generation instruction instructs the main strategy execution module and the sub-strategy execution module to select target datasets from the data model according to the data requirement labels. A complete decision model is established based on the main strategy model and its corresponding first target dataset, and the sub-strategy models and their corresponding second target datasets. Therefore, in this embodiment, the intelligent decision engine decouples the data model and the strategy model, enabling data model sharing between the main strategy model and the sub-strategy models; it automatically derives the input / output parameters of the sub-strategies, reducing manual configuration workload and lowering the error rate. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of the structure of an intelligent decision-making system provided in an embodiment of this application;

[0013] Figure 2 This is a schematic diagram of the structure of an intelligent decision engine provided in an embodiment of this application;

[0014] Figure 3 A schematic block diagram of another intelligent decision-making system provided in the embodiments of this application;

[0015] Figure 4 A flowchart illustrating a decision model generation method based on a data model and a strategy model, provided for an embodiment of this application;

[0016] Figure 5 A schematic diagram of a display page for intelligent policy configuration provided in an embodiment of this application;

[0017] Figure 6 A schematic diagram of the display page of the large language model dialogue page provided in the embodiments of this application;

[0018] Figure 7 A schematic diagram of the display page of the intelligent logic view provided in the embodiments of this application;

[0019] Figure 8 A schematic diagram illustrating the switching of conditional logic rules provided in the embodiments of this application;

[0020] Figure 9 A schematic diagram illustrating the configuration process of the data model provided in this application embodiment;

[0021] Figure 10 A schematic diagram illustrating the feedback process for modifying sub-strategy model content, provided in an embodiment of this application;

[0022] Figure 11 A flowchart illustrating the process of adding data to the sub-strategy model provided in this application embodiment;

[0023] Figure 12 A flowchart illustrating a decision-making method based on a data model and a strategy model, provided for an embodiment of this application;

[0024] Figure 13 A schematic diagram of the feedback process for modifying another sub-strategy model provided in this application embodiment;

[0025] Figure 14 A functional unit structure block diagram of a decision model generation device based on a data model and a strategy model provided in an embodiment of this application;

[0026] Figure 15 This is a schematic diagram of a data model deployment page provided in an embodiment of this application. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0028] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but in some embodiments includes steps or units not listed, or in some embodiments includes other steps or units inherent to these processes, methods, products, or apparatuses.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] In the embodiments of this application, "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent the following three situations: A exists alone; A and B exist simultaneously; B exists alone. Among them, A and B can be singular or plural.

[0031] In this embodiment, the symbol " / " can indicate that the preceding and following objects are in an "or" relationship. Alternatively, the symbol " / " can also represent a division sign, i.e., performing a division operation. For example, A / B can mean A divided by B.

[0032] In the embodiments of this application, "at least one item" or its similar expression refers to any combination of these items, including any combination of a single item or a plurality of items. "One or more" means one or more, while "multiple" means two or more. For example, "at least one item" of a, b, or c can represent the following seven cases: a, b, c; a and b; a and c; b and c; a, b, and c. Each of a, b, and c can be an element or a set containing one or more elements.

[0033] In the embodiments of this application, "equal to" can be used with "greater than" and is applicable to technical solutions used when "greater than" is used; it can also be used with "less than" and is applicable to technical solutions used when "less than" is used. When "equal to" is used with "greater than", it is not used with "less than"; when "equal to" is used with "less than", it is not used with "greater than".

[0034] In this application, the strategy model is a set of rules used to implement business decision-making logic, mainly defining the decision-making process, conditions, rules, and objectives. It is typically divided into a main strategy model (global decision-making framework) and sub-strategy models (local, segmented decision-making logic), forming a hierarchical decision-making system. The main strategy model is the top-level framework for decision-making, responsible for coordinating the overall decision-making logic and defining the overall decision-making objectives and processes. Sub-strategy models are sub-modules of the main strategy model, responsible for implementing single-dimensional decision-making logic and supporting the specific execution of the main strategy. For example, in a risk control scenario, the main strategy model might define the overall process of "whether to approve a loan" (e.g., first passing anti-fraud verification, then credit scoring, and finally combining repayment ability assessment), while the anti-fraud sub-strategy model (determining whether a user poses fraud risk), the credit scoring sub-strategy model (calculating the user's credit score), and the repayment ability sub-strategy model (assessing the user's repayment ability) are all sub-strategies of the risk control main strategy.

[0035] A data model is a structured definition of the data required for decision-making. It describes all data fields, formats, sources, and relationships involved in the decision-making process and serves as the "raw material" for the strategy model to operate. The goal is to isolate the strategy's parameter range, parameter validation logic, and the strategy's own configuration (internal rules, output parameter binding, etc.), managing these separately and enabling data reuse throughout the entire strategy's lifecycle. For example, it defines various types of business data, such as basic user information including user ID, age, gender, and occupation; transaction data including transaction amount, transaction time, and transaction location; and risk-related data including device fingerprints, credit scores, and overdue records.

[0036] The Decision Engine is an intelligent strategy decision-making platform designed for financial-grade scenarios. Based on a distributed architecture and a low-code visual interface, it provides enterprises with the ability to manage complete lifecycle business rules. By decoupling business logic from system code, it empowers non-technical personnel to quickly build complex decision flows using a strategy tree approach, achieving millisecond-level automated responses across multiple business scenarios such as risk control, approval, operations, and marketing, thus helping enterprises build a digital decision-making hub.

[0037] In traditional solutions, sub-strategies require manual declaration of input and output parameters, which is prone to omissions or errors. This application automatically derives the parameters required by the sub-strategy model through a two-way binding between the data model and the strategy model, enabling parameter reuse in the data model and reducing maintenance costs. Simultaneously, it achieves automatic derivation of the input / output parameters of the sub-strategy model; in special scenarios, only the differing parameters need to be manually adjusted, reducing the workload of manual definition.

[0038] In traditional solutions, sub-policies need to redefine existing parameters of the parent policy, leading to redundancy and synchronization errors. This application uses a global data model to allow the main policy model and sub-policy models to share the same set of parameter definitions, enabling changes to be made in one place to take effect globally.

[0039] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0040] Please see Figure 1 , Figure 1 This is a schematic diagram of the structure of an intelligent decision-making system provided in an embodiment of this application, as shown below. Figure 1 As shown, the intelligent decision-making system 1 includes an intelligent decision-making engine 10, a terminal device 20, and a strategy execution module 30. The intelligent decision-making engine 10, the terminal device 20, and the strategy execution module 30 are all interconnected.

[0041] The intelligent decision engine 10 is used to acquire data models and strategy models, and to issue execution instructions to the strategy execution module 30, instructing the strategy execution module 30 to acquire the target dataset and generate a decision model based on the strategy model and the target dataset.

[0042] In a specific implementation, the intelligent decision engine 10 can be an electronic device. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0043] See Figure 2 , Figure 2 This is a schematic diagram of the structure of an intelligent decision engine provided in an embodiment of this application, as shown below. Figure 2 As shown, the intelligent decision engine 10 includes at least one processor 21 and a memory, such as a read-only memory 22 and a random access memory 23, communicatively connected to the at least one processor 21. The memory stores computer programs executable by the at least one processor 21. The processor 21 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 22 or loaded from the storage unit 28 into the random access memory 23. The random access memory 23 can also store various programs and data required for the operation of the intelligent decision engine 10. The processor 21, read-only memory 22, and random access memory 23 are interconnected via a bus 24. An input / output interface 25 is also connected to the bus 24.

[0044] Multiple components in the intelligent decision engine 10 are connected to an interface, including: an input unit 26, such as a keyboard or mouse; an output unit 27, such as various types of displays or speakers; a storage unit 28, such as a disk or optical disc; and a communication unit 29, such as a network interface card (NIC), modem, or wireless transceiver. The communication unit 29 allows the electronic device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0045] Processor 21 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 21 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 21 performs the various methods and processes described above, such as a decision model generation method based on a data model and a policy model.

[0046] Specifically, terminal device 20 may include a front-end device applied to the user side, capable of data collection, data transmission, and other functions. This can be user equipment (UE) such as mobile phones, smartphones, laptops, digital radio receivers, personal digital assistants (PDAs), and tablet computers (PADs), handheld devices, in-vehicle devices, wearable devices, computing devices, or other processing devices connected to a wireless modem, mobile stations (MS), or mobile terminals. Alternatively, terminal device 20 may also be a software application capable of running on the aforementioned electronic devices. For example, it could be an app running on a mobile phone.

[0047] The strategy execution module 30 includes a main strategy execution module 31 and a sub-strategy execution module 32. The main strategy execution module 31 includes a main strategy processor 311 (such as a server CPU core or a dedicated computing chip). The main strategy execution module 31 obtains the original data model from the intelligent decision engine 10, executes the main strategy logic, takes the dataset in the original data model that meets the requirements of the main strategy as the first target dataset, and returns the first target dataset to the intelligent decision engine 10. The sub-strategy execution module 32 includes a sub-strategy processor 321 (such as a server CPU core or a dedicated computing chip), obtains a replica data model from the main strategy execution module 31, and uses the sub-strategy processor 321 to trim the replica data model as needed (such as filtering and deleting fields according to the data requirements of the sub-strategy) to obtain a second target dataset, and sends the second target dataset to the intelligent decision engine 10.

[0048] The main strategy execution module 31 and the sub-strategy execution module 32 exchange data through a strategy interaction interface.

[0049] In some embodiments, the sub-policy execution module 32 modifies the parameters of the replica data model according to its data requirements, and sends the modified content to the main policy execution module 31; the main policy execution module 31 synchronizes the modified content to the original data model.

[0050] In some embodiments, the sub-policy execution module 32 receives new user data from the terminal device 20 and sends the new user data to the main policy execution module 31; the main policy execution module 31 adds the new user data to the original data model.

[0051] In some embodiments, the sub-policy execution module 32 sends a data call request to the main policy execution module 31, which can call data from the original data model from the main policy execution module 31.

[0052] Please see Figure 3 , Figure 3 A schematic block diagram of another intelligent decision-making system provided in the embodiments of this application is shown below. Figure 3 As shown, the intelligent decision-making system 1 includes an output layer 11, an engine core layer 12, and a strategy execution layer 13. The engine core layer 12 is used to generate a decision model based on the data model and the strategy model. The strategy execution layer 13 is used to store the strategy model and manage strategy logic, variable parameter definitions, data filtering, etc., and supports version management and query calls for the main strategy and sub-strategies. The output layer 11 is used to store the decision model generated by the engine core layer 12 and output the decision results based on the decision model and actual business data.

[0053] Based on the above hardware structure, this application proposes a decision model generation method based on a data model and a strategy model.

[0054] Please see Figure 4 , Figure 4 A flowchart illustrating a decision model generation method based on a data model and a strategy model, provided in this application embodiment, is shown below. Figure 4 As shown, the method is applied to, for example Figure 1 The intelligent decision engine shown includes the following steps S401-S404:

[0055] Step S401: Obtain the strategy model.

[0056] The strategy model includes a main strategy model and sub-strategy models, and each strategy model contains data requirement tags.

[0057] Step S402: Obtain the data model.

[0058] The data model is associated with the main strategy model. The data model includes multiple datasets, each dataset has a unique data label, and the data label has a mapping relationship with the data requirement label.

[0059] Each strategy model (including the main strategy model and sub-strategy models) predefines the data types or field identifiers required for its operation during the design process. These identifiers are called "data requirement tags".

[0060] The data model that matches the current strategy model is primarily selected based on the data requirement tags of the main strategy model. In other words, there is a mapping relationship between the main strategy model and the data model.

[0061] Each data model comprises multiple datasets, each containing at least one variable parameter. Each dataset is assigned a unique data label, and these data labels are mapped to data requirement labels. In other words, each sub-strategy model within the strategy model can match its data requirement label with the data labels of the datasets within the data model, without needing to traverse the global dataset, thus ensuring that each sub-strategy model obtains its corresponding dataset.

[0062] For example, the mapping relationship between the strategy model and the data model is shown in Table 1 below. The main strategy model of the strategy model is the bank credit card approval strategy, and its sub-strategy models include identity verification sub-strategy, credit risk assessment sub-strategy, and repayment ability assessment sub-strategy. The data model corresponding to the main strategy model is the bank's core database. The bank's core database contains multiple datasets, such as personal identity information dataset (person_id_info), central bank credit dataset (central_credit), income certificate dataset (income_certificate), and loan behavior dataset (loan_behavior). Each dataset also includes specific variable parameters and labels.

[0063] The data requirement labels for the identity verification sub-model strategy include person_id_info, the data requirement labels for the credit risk assessment sub-model include central_credit and loan_behavior, and the data requirement labels for the repayment ability assessment sub-model include income_certificate and loan_behavior.

[0064] Furthermore, the strategy model and data model can be generated by the user in the terminal device, as follows:

[0065] The strategy model is generated as follows:

[0066] In some embodiments, a large language model can be used to assist in generating decision logic and the required variable parameters, helping users to achieve efficient and accurate strategy configuration.

[0067] For specific implementation details, please refer to [link / reference]. Figure 5 , Figure 5 This is a schematic diagram of a display page for intelligent policy configuration provided in an embodiment of this application, such as... Figure 5 As shown, the intelligent strategy configuration page 5 includes a strategy canvas 51 and a strategy node display bar 52. The strategy node display bar 52 is used to display different process nodes, including but not limited to start nodes, condition judgment nodes, branch execution nodes, and end nodes. Users can drag different process nodes onto the strategy canvas 51 to add nodes and form a strategy tree (e.g., ...). Figure 5 The strategy tree example in the strategy canvas 51 includes a start node 1, a condition judgment node 1, a branch execution node 1, a branch execution node 2, a branch execution node 3, a branch execution node 4, a branch execution node 5, a branch execution node 6, an end node 1, an end node 2, and an end node 3. In addition, the strategy node display bar 52 also includes a strategy engine assistant control 521. When the user clicks the strategy engine assistant control 521, the large language model dialogue page 6 is displayed on the intelligent strategy configuration page 5. Figure 6 , Figure 6 This is a schematic diagram of the display page of the large language model dialogue page provided in the embodiment of this application. Users can enter questions to query on the large language model dialogue page 6, or use the large language model to assist in generating policy logic and the datasets needed, and instruct the large language model to directly generate a policy tree based on the generated policy logic.

[0068] In some embodiments, policy logic is configured using a smart logic view, see [link to configuration]. Figure 7 , Figure 7 This is a schematic diagram of the display page of the intelligent logic view provided in the embodiments of this application, such as... Figure 7 As shown, page 7 of the Smart Logic View uses nested groups to construct multi-level conditional relationships, connecting different conditional branches through logical operators such as "AND" and "OR". For example... Figure 7 There are multiple nested groups, and specific condition judgments can be set within each nested group.

[0069] By switching the logical view 71 control, you can Figure 7 The conditional logic rule configuration set in the middle is switched to Figure 8 The conditional thinking guide shown Figure 8Users can better view and manage complex nested conditions, avoiding logical confusion and making rule maintenance and understanding easier. Compared to traditional code-based logic writing, this visual conditional logic configuration method reduces the technical knowledge requirements for business personnel. Complex business logic can be defined without writing code, through simple parameter selection and condition configuration, improving the efficiency and accuracy of rule configuration.

[0070] The data model is generated using the following methods:

[0071] Users configure the specific details of each variable parameter on the data model deployment page, based on the dataset required by the strategy model. Please refer to [link / reference]. Figure 15 , Figure 15 This is a schematic diagram of a data model deployment page provided in an embodiment of this application, such as... Figure 15 As shown, the data model deployment page 15 is used to set variable codes, variable names, variable types (such as string, long integer, int, float, etc.), variable states (including enabled and disabled states), variable tags (such as risk control, approval, etc.), variable categories, variable read / write attributes (including all strategies are readable, all strategies are writable, and some strategies are writable), validation rules, and remarks. Validation rules are used to verify whether data types, data ranges, etc., conform to preset ranges. For example, if the normal range for age is set to 0-200, an error message will be automatically displayed if the user enters an age range exceeding 0-200. When the read / write attribute is "some strategies are writable," the strategy code for the partially writable strategy must be entered. The data model deployment page 15 also includes an add enumeration value control. Clicking this control allows users to enter enumeration values, their meanings, and sorting. The added enumeration values ​​for this variable are displayed on the data model deployment page 15. Editing or deleting operations can be performed on the added enumeration values. After setting, clicking the submit control generates the data model for that variable, or clicking the cancel control stops the data model generation task.

[0072] The variable classification is associated with the dropdown through the dictionary encoding dataSetVarCategory, where the dictionary value of the engine's built-in strategy process variable is stgTempType.

[0073] One variable can correspond to multiple variable labels. When configuring a strategy, variables can be filtered based on variable labels. For example, a risk control strategy can filter variables that contain risk control labels.

[0074] The variable type cannot be changed.

[0075] Table 1 shows the mapping relationship between the strategy model and the data model provided in the embodiments of this application.

[0076]

[0077] Step S403: Input the data model into the strategy model and send a decision model generation instruction to the strategy execution module.

[0078] The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to select the target dataset from the data model according to the data requirement labels.

[0079] In some embodiments, the specific steps for the main strategy execution module and the sub-strategy execution module to select the target dataset from the data model according to the data requirement labels include: the main strategy execution module uses the original data model as the first target dataset; the main strategy execution module copies the original data model to obtain multiple replica data models, and inputs the replica data models into the sub-strategy execution module; the sub-strategy execution module trims and / or modifies the replica data models according to the data requirement labels to obtain the second target dataset.

[0080] The intelligent decision engine uses a data model to associate content with the main strategy model, and the main strategy execution module directly uses the original data model as the first target dataset. Specifically, if the main strategy model is a bank credit card approval strategy, the first target dataset includes the bank's core database, specifically including datasets such as personal identity information sets, central bank credit information datasets, income verification datasets, and lending behavior datasets.

[0081] The main strategy execution module copies the acquired original data model to obtain multiple replica data models. These replica data models are then distributed to each sub-strategy execution module. Each sub-strategy execution module filters the datasets required by its own sub-strategy model based on the data requirement tags and the data tags of the datasets in the data model. Unnecessary datasets are removed. If the sub-strategy model has special parameter settings, the variable parameters in the original data model can be modified. Finally, the filtered and modified dataset is determined as the second target dataset.

[0082] As can be seen, in this embodiment, the main strategy model and the sub-strategy model share the same data model. Once the data model is defined, the parameters are available everywhere within the strategy model, achieving parameter reuse and reducing maintenance costs. Simultaneously, it enables automatic derivation of the input / output parameters of the sub-strategy model; in special scenarios, only manual adjustment of the differing parameters is required, reducing the workload of manual definition.

[0083] For specific implementation details, see Figure 9 , Figure 9 This is a schematic diagram of the configuration process of the data model provided in the embodiments of this application, such as... Figure 9As shown, the strategy model includes a main strategy model and multiple sub-strategy models (such as sub-strategy model 1.1, sub-strategy model 1.2, sub-strategy model 1.3, sub-strategy model 2.1, sub-strategy model 2.2, sub-strategy model 2.3, and sub-strategy model 2.4). The main strategy model selects the original data model and replicates it into multiple replica data models. The main strategy model determines the original data model as the first target dataset and distributes the replica data models to the sub-strategy models. The sub-strategy models trim and / or modify the replica data models according to their own data requirement labels. Finally, each sub-strategy model obtains its corresponding second target dataset.

[0084] In this application, under normal circumstances, without risk operation expansion, the sub-strategy model can usually only call the dataset with the same data requirement label. The advantage of doing so is that it can match data that is completely aligned with the sub-strategy model to a limited extent, reducing the verification cost of complex associations.

[0085] After the risk operations expansion, the system grants permissions, allowing sub-strategy models to access public datasets. For example, without the risk operations expansion, the credit risk assessment sub-strategy of the main credit card approval strategy could only use the central bank's credit reporting dataset and lending behavior dataset. After the system grants access to public datasets, the identity verification sub-strategy can access the "industry fraud list" from the public dataset, supplementing the judgment dimensions of personal credit risk assessment. This approach both restricts arbitrary cross-dataset access (preventing data abuse) and allows businesses to expand public data as needed (ensuring flexibility).

[0086] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset, the method further includes: the sub-policy execution module sending the modified content of the replica data model to the main policy execution module; the main policy execution module modifying the original data model according to the modified content.

[0087] For specific implementation details, see Figure 10 , Figure 10 This is a schematic diagram of the feedback process for modifying the content of a sub-strategy model provided in an embodiment of this application. After the sub-strategy execution module modifies the variable parameters in the replica data model, it needs to feed back the modified content to the main strategy execution module. The main strategy execution module directly modifies the original data model to achieve global modification.

[0088] Please see Figure 13 , Figure 13 This is a schematic diagram illustrating the feedback process for modifying another sub-strategy model provided in an embodiment of this application, as shown below. Figure 13As shown, the original data model and the replica data model are input into the main strategy model and the sub-strategy model, respectively. The main strategy model saves the original data model, and the sub-strategy model trims / modifies the replica data model according to the data requirement tags, outputs the modified content, and overwrites the corresponding content in the original data model.

[0089] Step S404: Establish a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

[0090] As can be seen, in this embodiment, the intelligent decision engine acquires a strategy model and a data model. The strategy model includes a main strategy model and sub-strategy models. Each strategy model contains data requirement labels. The data model is associated with the main strategy model and includes multiple datasets. Each dataset has a uniquely identified data label, and there is a mapping relationship between the data labels and the data requirement labels. The data model is then input into the strategy model, and a decision model generation instruction is sent to the strategy execution module. The decision model generation instruction instructs the main strategy execution module and the sub-strategy execution module to select target datasets from the data model according to the data requirement labels. A complete decision model is established based on the main strategy model and its corresponding first target dataset, and the sub-strategy models and their corresponding second target datasets. Therefore, in this embodiment, the intelligent decision engine decouples the data model and the strategy model, enabling data model sharing between the main strategy model and the sub-strategy models; it automatically derives the input / output parameters of the sub-strategies, reducing manual configuration workload and lowering the error rate.

[0091] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement label to obtain the second target dataset, the method further includes: the sub-policy execution module receiving new user data from the terminal device; the sub-policy execution module sending the new user data to the main policy execution module; and the main policy execution module saving the new user data and its source to the original data model.

[0092] When new data is added within a single sub-strategy model, the sub-strategy model needs to synchronize the new user data to the original dataset of the main strategy model. After synchronization to the original data model, other sub-strategy models can also access the new user data.

[0093] For specific implementation details, see Figure 11 , Figure 11 This is a flowchart illustrating the process of adding data to the sub-strategy model provided in the embodiments of this application, such as... Figure 11As shown, the first sub-strategy model includes multiple subordinate strategy models. The second target dataset of the first sub-strategy model is the credit granting dataset, which includes variable parameters varA, varB, and varC. The second sub-strategy model includes multiple subordinate strategy models. The second target dataset of the second sub-strategy model is the credit adjustment dataset, which includes variable parameters varA, varB, and varD. The original data model includes varA, varB, varC, and varD. The sub-strategy models can achieve cross-business domain calls.

[0094] A new parameter varE is added to the second sub-strategy model. The sub-strategy execution module sends the new parameter varE to the original data model. The first and second sub-strategy models can call the new parameter varE from the original data model and mark its source.

[0095] When calling variable parameters, the configured business validation logic is brought over by default. Users can edit and modify the business validation logic, and the specific business validation logic configured in the strategy will prevail when the strategy is executed.

[0096] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement label to obtain the second target dataset, the method further includes: the main policy execution module receiving a data call request from the sub-policy execution module; and the main policy execution module sending the corresponding data to the sub-policy execution module according to the data call request.

[0097] The sub-strategy model periodically accesses the main strategy model. After the original data model adds or modifies variable parameters, the sub-strategy execution module can send a data call request to the main strategy execution module to obtain the added or modified variable parameters.

[0098] In some embodiments, after establishing a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset, the method further includes: receiving strategy data from a terminal device, the strategy data including real business data and simulated business data, the simulated business data referring to virtual business data generated based on a large language model; inputting the strategy data into the decision model, calculating the contribution rate of variable parameters in each data model, the contribution rate referring to the probability that changes in variable parameters lead to changes in the decision result; detecting that the contribution rate of a variable parameter exceeds a preset normal value range; and optimizing the strategy model corresponding to the variable parameter.

[0099] The intelligent decision engine generates a decision model and then simulates the decision outcome by inputting strategy data. This decision data includes both real and virtual business data, and the weighting of these two types of data can be dynamically adjusted (e.g., "90% real + 10% simulated") to accelerate strategy iteration while ensuring system stability.

[0100] Virtual business data can be intelligently generated through large language models.

[0101] The contribution rate calculation involves manually changing the value of each variable (e.g., user age automatically) while keeping other variables constant, and observing the changes in the decision outcome. The more frequent the changes, the greater the variable's influence on the decision, and the higher the contribution rate. This "influence frequency" is then converted into a percentage (e.g., user age automatically influenced 50,000 decision changes, with a total of 100,000 simulations, so 50%). Essentially, this method uses a "controlled variable method" to conduct large-scale simulations, statistically analyzing the frequency with which each variable "causes a change in decision." The higher the frequency, the higher the contribution rate.

[0102] When the contribution rate does not meet expectations (too high or too low), adjustments can be made to the strategy logic and improve data quality. Specifically:

[0103] When the contribution rate of a variable parameter is too high, the single judgment rule can be modified into a multi-variable joint judgment to reduce the decision influence of a single variable and make the result more balanced.

[0104] When the contribution rate of a variable parameter is too low, data quality should be checked and data validation should be increased. For example, if the error rate of the variable parameter "name" is high, it may be ignored during decision-making. Validation can be increased by comparing the entered name with the name on the identification document in real time to correct errors, thereby improving data quality and ensuring that the variable parameter plays a normal role in decision-making.

[0105] Please see Figure 12 , Figure 12 A flowchart illustrating a decision-making method based on a data model and a strategy model, provided in an embodiment of this application, is shown below. Figure 12 As shown, the method is applied to an intelligent decision engine in an intelligent decision system. The intelligent decision system includes the intelligent decision engine and a strategy execution module. The strategy execution module includes a main strategy execution module and a sub-strategy execution module. The method includes the following steps S1201-S1207:

[0106] Step S1201: Receive a decision request from the terminal device.

[0107] Step S1202: Obtain the strategy model according to the decision request.

[0108] The strategy model includes a main strategy model and sub-strategy models, and each strategy model contains data requirement tags.

[0109] Step S1203: Obtain the data model based on the main strategy model.

[0110] The data model is associated with the main strategy model. The data model includes multiple datasets, each dataset has a unique data label, and the data label has a mapping relationship with the data requirement label.

[0111] Step S1204: Input the data model into the strategy model and send a decision model generation instruction to the strategy execution module.

[0112] The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to select the target dataset from the data model according to the data requirement labels.

[0113] Step S1205: Establish a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

[0114] Step S1206: Receive policy data from the terminal device.

[0115] Step S1207: Input the strategy data into the decision model to obtain the decision result.

[0116] Once the intelligent decision engine generates a decision model, actual business data can be input into the decision model to obtain the decision results.

[0117] The above primarily describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the server includes the corresponding hardware structure and / or software modules for executing each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0118] This application embodiment can divide the server into functional units according to the above method example. For example, each function can be divided into different functional units, or two or more functions can be integrated into one processing module. The integrated unit can be implemented in hardware or as a software program module. It should be noted that the unit division in this application embodiment is illustrative and only represents a logical functional division, while other division methods may be used in actual implementation.

[0119] In the case of using integrated units, please refer to Figure 14 , Figure 14 A functional unit structure diagram of a decision model generation device based on a data model and a strategy model provided in this application embodiment. The decision model generation device 14 based on a data model and a strategy model includes:

[0120] The acquisition unit 1401 is used to acquire a strategy model, which includes a main strategy model and sub-strategy models, each strategy model containing a data requirement label; and to acquire a data model, which is associated with the main strategy model, the data model including multiple datasets, each dataset having a uniquely identified data label, and the data label having a mapping relationship with the data requirement label.

[0121] Processing unit 1402 is configured to input the data model into the strategy model and send a decision model generation instruction to the strategy execution module. The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to select target datasets from the data model according to data requirement labels. It also establishes a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

[0122] As can be seen, in this embodiment, the intelligent decision engine acquires a strategy model and a data model. The strategy model includes a main strategy model and sub-strategy models. Each strategy model contains data requirement labels. The data model is associated with the main strategy model and includes multiple datasets. Each dataset has a uniquely identified data label, and there is a mapping relationship between the data labels and the data requirement labels. The data model is then input into the strategy model, and a decision model generation instruction is sent to the strategy execution module. The decision model generation instruction instructs the main strategy execution module and the sub-strategy execution module to select target datasets from the data model according to the data requirement labels. A complete decision model is established based on the main strategy model and its corresponding first target dataset, and the sub-strategy models and their corresponding second target datasets. Therefore, in this embodiment, the intelligent decision engine decouples the data model and the strategy model, enabling data model sharing between the main strategy model and the sub-strategy models; it automatically derives the input / output parameters of the sub-strategies, reducing manual configuration workload and lowering the error rate.

[0123] In some embodiments, the specific steps for the main strategy execution module and the sub-strategy execution module to select the target dataset from the data model according to the data requirement labels include: the main strategy execution module uses the original data model as the first target dataset; the main strategy execution module copies the original data model to obtain multiple replica data models, and inputs the replica data models into the sub-strategy execution module; the sub-strategy execution module trims and / or modifies the replica data models according to the data requirement labels to obtain the second target dataset.

[0124] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset, the method further includes: the sub-policy execution module sending the modified content of the replica data model to the main policy execution module; and the main policy execution module modifying the original data model according to the modified content.

[0125] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement label to obtain the second target dataset, the method further includes: the sub-policy execution module receiving new user data from the terminal device; the sub-policy execution module sending the new user data to the main policy execution module; and the main policy execution module saving the new user data and its source to the original data model.

[0126] In some embodiments, after the sub-policy execution module trims and / or modifies the replica data model according to the data requirement label to obtain the second target dataset, the method further includes: the main policy execution module receiving a data call request from the sub-policy execution module; and the main policy execution module sending the corresponding data to the sub-policy execution module according to the data call request.

[0127] In some embodiments, after the processing unit 1402 establishes a complete decision model based on the main strategy model and its corresponding first target dataset, and the sub-strategy model and its corresponding second target dataset, the processing unit 1402 is further configured to receive strategy data from a terminal device. The strategy data includes real business data and simulated business data, where simulated business data refers to virtual business data generated based on a large language model. The processing unit 1402 inputs the strategy data into the decision model, calculates the contribution rate of variable parameters in each data model, where the contribution rate refers to the probability that a change in the variable parameter will lead to a change in the decision result. The processing unit 1402 detects that the contribution rate of a variable parameter exceeds a preset normal value range and optimizes the strategy model corresponding to that variable parameter.

[0128] This application provides a computer-readable storage medium storing a computer program / instructions thereon, which, when executed by a processor, implement the steps of any possible embodiment of the method.

[0129] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0130] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0131] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0132] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0133] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0134] If the integrated units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0135] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0136] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for generating decision models based on data models and strategy models, characterized in that, An intelligent decision engine applied in an intelligent decision-making system, the intelligent decision-making system including the intelligent decision engine and a strategy execution module, the strategy execution module including a main strategy execution module and a sub-strategy execution module, the method including: Obtain the strategy model, which includes a main strategy model and sub-strategy models, and each strategy model contains data requirement tags; Obtain a data model, which is associated with the main strategy model. The data model includes multiple datasets, each dataset is assigned a unique data label, and the data label is mapped to the data requirement label. The data model is input into the strategy model, and a decision model generation instruction is sent to the strategy execution module. The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to filter out the target dataset from the data model according to the data requirement labels. A complete decision-making model is established based on the main strategy model and its corresponding first target dataset, and the sub-strategy model and its corresponding second target dataset.

2. The method according to claim 1, characterized in that, The specific steps for the main strategy execution module and the sub-strategy execution module to select the target dataset from the data model according to the data requirement labels include: The main strategy execution module uses the original data model as the first target dataset; The main strategy execution module copies the original data model to obtain multiple replica data models, and inputs the replica data models into the sub-strategy execution module; The sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset.

3. The method according to claim 2, characterized in that, After the sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset, the method further includes: The sub-policy execution module sends the modifications to the replica data model to the main policy execution module; The main strategy execution module modifies the original data model according to the modified content.

4. The method according to claim 2, characterized in that, After the sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset, the method further includes: The sub-policy execution module receives new user data from the terminal device; The sub-policy execution module sends the newly added user data to the main policy execution module; The main strategy execution module saves the newly added user data and its source to the original data model.

5. The method according to claim 2, characterized in that, After the sub-policy execution module trims and / or modifies the replica data model according to the data requirement tags to obtain the second target dataset, the method further includes: The main strategy execution module receives a data call request from the sub-strategy execution module; The main strategy execution module sends the corresponding data to the sub-strategy execution module according to the data call request.

6. The method according to claim 1, characterized in that, After establishing a complete decision model based on the main strategy model and its corresponding first target dataset, and the sub-strategy model and its corresponding second target dataset, the method further includes: Receive policy data from terminal devices. The policy data includes real business data and simulated business data. Simulated business data refers to virtual business data generated based on a large language model. The strategy data is input into the decision model, and the contribution rate of the variable parameters in each data model is calculated. The contribution rate refers to the probability that a change in the variable parameters will lead to a change in the decision result. The contribution rate of the variable parameter was detected to exceed the preset normal value range; Optimize the strategy model corresponding to the variable parameter.

7. A decision-making method based on data models and strategy models, characterized in that, An intelligent decision engine applied in an intelligent decision-making system, the intelligent decision-making system including the intelligent decision engine and a strategy execution module, the strategy execution module including a main strategy execution module and a sub-strategy execution module, the method including: Receive decision requests from terminal devices; Based on the decision request, a strategy model is obtained, which includes a main strategy model and sub-strategy models, and each strategy model contains data requirement tags. A data model is obtained based on the main strategy model. The data model is associated with the main strategy model. The data model includes multiple datasets. Each dataset is assigned a unique data label. The data label is mapped to the data requirement label. The data model is input into the strategy model, and a decision model generation instruction is sent to the strategy execution module. The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to filter out the target dataset from the data model according to the data requirement labels. A complete decision-making model is established based on the main strategy model and its corresponding first target dataset, and the sub-strategy model and its corresponding second target dataset. Receive policy data from the terminal device; The strategy data is input into the decision model to obtain the decision results.

8. A device for generating a decision model based on a data model and a strategy model, characterized in that, The device, applied to an intelligent decision-making engine, includes: An acquisition unit is used to acquire a strategy model, which includes a main strategy model and sub-strategy models, each strategy model containing a data requirement label; and to acquire a data model, which is associated with the main strategy model, the data model including multiple datasets, each dataset having a uniquely identified data label, and the data label having a mapping relationship with the data requirement label. The processing unit is configured to input the data model into the strategy model and send a decision model generation instruction to the strategy execution module. The decision model generation instruction is used to instruct the main strategy execution module and the sub-strategy execution module to select target datasets from the data model according to data requirement labels; and to establish a complete decision model based on the main strategy model and its corresponding first target dataset, the sub-strategy model and its corresponding second target dataset.

9. An intelligent decision-making engine, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the step instructions of the method as described in any one of claims 1 to 6 when it invokes the computer program in the memory.

10. An intelligent decision-making system, characterized in that, It includes an intelligent decision engine, a terminal device, and a strategy execution module. The strategy execution module includes a main strategy execution module and a sub-strategy execution module. The intelligent decision engine is used to execute the step instructions in the method as described in any one of claims 1 to 6.

Images