A decision model generation method and apparatus

Abstract

The embodiment of the application provides a decision model generation method and device, which can be used in the field of artificial intelligence technology, and the method comprises the following steps: training a meta-explanation learning framework through pre-acquired data samples and background knowledge to obtain a decision model generation model; and generating a decision model according to acquired operation operations and data types through the decision model generation model, so that the flexibility of a formula editor can be increased, the office efficiency of a user can be improved, and the user experience can be improved.

Description

Technical Field

[0001] This invention relates to the field of computer data processing technology, particularly to the field of artificial intelligence technology, and especially to a method and apparatus for generating decision models. Background Technology

[0002] Currently, when using the anti-fraud decision engine system, the formula editor is often cumbersome to use. Users obtain calculation formulas by dragging or double-clicking, but each operation can only perform one calculation, and the result of the previous calculation cannot be directly used as a parameter for subsequent calculations. This results in a lack of flexibility in the formula editor, which reduces the user experience and lowers user productivity. Summary of the Invention

[0003] One object of this invention is to provide a decision model generation method that increases the flexibility of formula editors, improves user efficiency, and enhances user experience. Another object of this invention is to provide a decision model generation apparatus. A further object of this invention is to provide a computer-readable medium. A still other object of this invention is to provide a computer device.

[0004] To achieve the above objectives, this invention discloses a method for generating a decision model, comprising:

[0005] By using pre-acquired data samples and background knowledge, the meta-explanation learning framework is trained to obtain a decision model generation model;

[0006] The decision model is generated by generating a model based on the acquired computational operations and data types.

[0007] Preferably, the background knowledge includes a feature information pool;

[0008] Before training the meta-explanation learning framework using pre-acquired data samples and background knowledge to obtain the decision model generator model, the following steps are also included:

[0009] Obtain a specified number of data samples, which include input samples and expected formulas. The input samples include at least one operation and a data type.

[0010] By specifying a logic programming language, a feature information pool is constructed based on preset functional relationships.

[0011] Preferably, the background knowledge also includes the selected meta-rules;

[0012] By using pre-acquired data samples and background knowledge, the meta-explanation learning framework is trained to obtain a decision model generation model, including:

[0013] The meta-interpretive learning framework is trained by inputting at least one operation, expectation formula, meta-rule, and feature information pool, and outputs a decision model to generate a model.

[0014] Preferably, a decision model is generated by generating a model based on the acquired computational operations and data types, including:

[0015] Input the computational operations and data types into the decision model to generate the model, and output the decision model.

[0016] Preferably, a meta-rule includes a meta-rule header and a corresponding meta-rule body;

[0017] The meta-interpretive learning framework is trained by inputting at least one operational operation, expectation formula, meta-rule, and feature information pool to output a decision model generation model, including:

[0018] Take at least one computational operation and its corresponding desired formula as the atomic target;

[0019] Match the atomic target with the meta-rule header to obtain the corresponding meta-rule header;

[0020] Match the meta-rule body corresponding to the meta-rule head with the restricted clauses in the feature information pool;

[0021] If the meta-rule body can match a restricted clause, save the restricted clause;

[0022] If the meta-rule body cannot match a restricted clause, the meta-rule body is used as a new atomic target, and the steps of matching the atomic target with the meta-rule head and matching the corresponding meta-rule head are repeated until the matching is completed.

[0023] Preferred options also include:

[0024] If the decision model generation model is not output within the set time limit, the data samples and background knowledge are adjusted, and the steps of training the meta-interpretive learning framework with the pre-acquired data samples and background knowledge are repeated to obtain the decision model generation model.

[0025] Preferably, the number of input samples is 5 to 10.

[0026] Preferably, after generating the decision model through the decision model generation process, based on the acquired computational operations and data types, the process further includes:

[0027] User data is input into the decision model for calculation to obtain the decision result;

[0028] The decision-making results are evaluated by setting evaluation rules, and the evaluation results are obtained.

[0029] The present invention also discloses a decision model generation device, comprising:

[0030] The training unit is used to train the meta-interpretive learning framework using pre-acquired data samples and background knowledge to obtain the decision model generation model.

[0031] The generation unit is used to generate a model from the decision model, based on the acquired computational operations and data types.

[0032] The present invention also discloses a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0033] The present invention also discloses a computer device, including a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, wherein the processor executes the program to implement the method described above.

[0034] This invention trains a meta-interpretive learning framework using pre-acquired data samples and background knowledge to obtain a decision model generation model. This decision model generation model, based on the acquired computational operations and data types, generates a decision model, increasing the flexibility of the formula editor, improving user efficiency, and enhancing the user experience. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 A flowchart illustrating a decision model generation method provided in an embodiment of the present invention;

[0037] Figure 2 A flowchart illustrating another decision model generation method provided in this embodiment of the invention;

[0038] Figure 3 This is a schematic diagram of the structure of a decision model generation device provided in an embodiment of the present invention;

[0039] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] It should be noted that the decision model generation method and apparatus disclosed in this application can be used in the field of artificial intelligence technology, or in any field other than artificial intelligence technology. The application field of the decision model generation method and apparatus disclosed in this application is not limited.

[0042] To facilitate understanding of the technical solution provided in this application, the relevant content of the technical solution will be described below. The meta-interpretive learning framework involved in this invention differs from common statistical-based big data learning. The meta-interpretive learning framework is a state-of-the-art method of inductive logic programming (ILP), belonging to few-shot learning. ILP provides hypotheses in symbolic form, which, compared to black-box models like neural networks, has excellent expressiveness and understandability. The meta-interpretive learning framework utilizes highly abstract logical languages ​​and functions to describe input objects, output objects, and background knowledge. Through inductive matching, it learns a set of programming in logic (Prolog) programs. These programs describe the rules for generating target data from source data. Reusing the learned Prolog programs can generate data that conforms to the same rules. Since the Prolog programs are visible, developers can optimize the code on top of the learned existing Prolog programs.

[0043] The meta-interpretive learning framework works as follows:

[0044] Given a set of samples ε and background knowledge β, input (β, ε) to a meta-interpretive learning framework. The framework either outputs a deterministic procedure H consistent with the input hypothesis, or terminates, declaring that no procedure can be found. Here, the given samples ε are the training data provided for learning. Let f(x, y) denote the learning objective, i.e., the final deterministic procedure H, where x is the input data and y is the output data. The background knowledge β consists of a restricted clause D and meta-rules M.

[0045] To address the aforementioned technical problems, a decision model generation method is provided based on the working principle of the meta-interpretive learning framework. The implementation process of the decision model generation method provided in this embodiment of the invention is illustrated below using a decision model generation device as the executing entity. It is understood that the executing entity of the decision model generation method provided in this embodiment of the invention includes, but is not limited to, a decision model generation device.

[0046] Figure 1 A flowchart of a decision model generation method provided in an embodiment of the present invention is shown below. Figure 1 As shown, the method includes:

[0047] Step 101: Train the meta-interpretive learning framework using pre-acquired data samples and background knowledge to obtain the decision model generation model.

[0048] Step 102: Generate a model through the decision model. Based on the acquired computational operations and data types, generate a decision model.

[0049] In the technical solution provided by the embodiments of the present invention, the meta-interpretive learning framework is trained by pre-acquired data samples and background knowledge to obtain a decision model generation model; the decision model generation model generates a decision model based on the acquired operation and data type, which can increase the flexibility of the formula editor, improve the user's office efficiency, and enhance the user experience.

[0050] Figure 2 A flowchart of another decision model generation method provided in an embodiment of the present invention is shown below. Figure 2 As shown, the method includes:

[0051] Step 201: Obtain a specified number of data samples, which include input samples and expected formulas. The input samples include at least one operation and data type.

[0052] In this embodiment of the invention, each step can be executed by a decision model generation device.

[0053] In this embodiment of the invention, a specified number of data samples need to be obtained before entering the learning process. The number of data samples can be set according to the actual situation, and this embodiment of the invention does not limit this. As an optional solution, the number of data samples is 5-10. The data samples include input samples and expected formulas. In this embodiment of the invention, input samples refer to at least one operation and data type that the user needs to use. The operation includes, but is not limited to, greater than, sum, maximum value, average value, and rounding operations. The data type includes, but is not limited to, decimal and integer. The expected formula in this embodiment of the invention refers to the formula that the user expects to output.

[0054] It is worth noting that when the input sample includes multiple operations, the operations must be performed in the order of the input samples. For example, if the input sample includes the average and rounding operations, the average of the data must be calculated first, and then the average must be rounded.

[0055] The expected formula refers to the formula corresponding to the expected input sample. For example, if the input sample includes three data points: data 1 is a decimal, data 2 is an integer, and data 3 is an integer, and the operations are averaging and rounding to two decimal places, then the expected formula is: [decimal] result = #rounding(#average([decimal] parameter1, [integer] parameter2, [integer] parameter3), 2), where the content in [] represents the data type, and the content of # represents the operation. It is worth noting that the expected formula can also have other styles, and the specific style of the expected formula can be set by the developers according to their needs. This embodiment of the invention does not limit this.

[0056] Step 202: Construct a feature information pool by using a specified logic programming language and according to preset operation parameters.

[0057] In this embodiment of the invention, background knowledge needs to be acquired before learning begins. This background knowledge includes a feature information pool, which contains multiple restricted clauses. The feature information pool is pre-constructed using a specified logic programming language based on preset operations. The specified logic programming language is Prolog, and the operations are all possible operations. That is, all possible operations are described using Prolog, and the operations described in Prolog are the restricted clauses. These restricted clauses are concentrated in the set feature information pool so that, during actual learning, the corresponding restricted clauses can be matched from the feature information pool based on data samples. The feature information pool includes multiple restricted clauses. For example, `contains_var(Sub,Term)` determines whether the sub-item (Sub) is contained in the main item (Term), returning true if successful. Another example is `ifthenelse(Term1,Term2,Term3)`, which executes Term2 if Term1 is true, otherwise executes Term3.

[0058] In this embodiment of the invention, the language logic of Prolog is very simple, leaning towards natural language, making it easy for researchers to learn and optimize the code.

[0059] Step 203: Train the meta-interpretive learning framework using data samples and background knowledge to obtain the decision model generation model.

[0060] In this embodiment of the invention, the background knowledge also includes the selected meta-rules. The specific selection of meta-rules needs to be made by developers based on actual business requirements. Common meta-rules are shown in Table 1:

[0061] Table 1

[0062] Meta-rule type Meta-rule expressions Step Instance P(A,B) True Identity P(A,B)←Q(A,B) P > Q Front-connected type P(A,B)←Q(A),R(A,B) P > Q, Q > R Rear connection type P(A,B)←Q(A,B),R(B) P > Q, Q > R chain P(A,B)←Q(A,C),R(C,B) P > Q, Q > R Tail recursion P(A,B)←Q(A,C),P(C,B) P > Q, Q > R, A > B > C Binary identity P(A,B)←Q(A,C),R(C,B) P > Q, Q > R

[0063] In this context, A, B, and C are variables, and P, Q, and R are predicates. Predicates are the functional relationships needed for the generative decision model. P(A, B) and Q(A, B) represent different relationships or actions between variables A and B. The order represents the priority, which is a priority relationship that the variables in the meta-rule need to satisfy. The arrow represents the implication relationship. The left part of the implication relationship is the meta-rule head, and the right part is the meta-rule body. For example, P(A, B) ← Q(A, B) means that if Q(A, B) is true, then P(A, B) is true. The implication of P(A, B) and Q(A, B) is denoted as P(A, B) ← Q(A, B). P(A, B) ← Q(A, C), P(C, B) means that if both Q(A, C) and P(C, B) are true, then P(A, B) is true.

[0064] Instance-type metarules P(A,B) have no implication relation, and therefore no order-based relational constraints. These metarules are often used to represent restricted clauses in background knowledge. The symbol > signifies proportionality; the variable on the left of > has higher priority than the variable on the right. For example, P>Q means that the predicate P must have higher priority than Q, specifically, the predicate P appears in the text before the predicate Q. This constraint can reduce search costs when searching for target programs.

[0065] Specifically, at least one operation, an expected formula, meta-rules, and a feature information pool are input into the meta-interpretive learning framework for training, outputting a decision model generation model. The training process involves recursively proving all given operations and their corresponding expected formulas sequentially. Specifically: at least one operation and its corresponding expected formula serve as atomic targets. The framework matches these atomic targets with meta-rule heads to obtain the corresponding meta-rule heads; then, it matches the meta-rule body corresponding to the meta-rule head with restricted clauses in the feature information pool. If the meta-rule body can match a restricted clause, the restricted clause is saved; otherwise, it becomes a new target and continues to be matched with the meta-rule head until a match is found. It is worth noting that the training of the decision model generation model needs to be completed within a specified time. If a Prolog program is not output within the specified time, the data samples and background knowledge need to be adjusted. Once all given operations and their corresponding expected formulas have been proven, a Prolog program consisting of multiple assigned meta-rules is obtained.

[0066] In this embodiment of the invention, the decision model generation model is based on few-shot learning. Users only need to provide the operation and data type to obtain the correct decision model, which improves user experience and office efficiency.

[0067] Step 204: Generate a model through the decision model. Based on the acquired computational operations and data types, generate a decision model.

[0068] Specifically, the operation and data type are input into the decision model to generate the model, and the decision model is output.

[0069] In this embodiment of the invention, the decision model generation model is the learned Prolog program. By reusing the learned Prolog program, the corresponding decision model can be obtained through the input operations and data types.

[0070] The formula editor within the engine operates independently of existing business processes, serving as a common decision-making tool suitable for data processing and computation needs across various scenarios. For example, the decision model can be applied to the anti-fraud field to comprehensively calculate anti-fraud assessment results from user data. Specifically, user data is input into the decision model for calculation to obtain a decision result; the decision result is then evaluated using set evaluation rules to obtain an assessment result. These evaluation rules can be set according to actual circumstances, and this embodiment of the invention does not impose limitations on them. As an optional approach, the evaluation rule is set such that if the decision result is greater than or equal to a decision threshold, an assessment result of passing the evaluation is generated, indicating no risk; if the decision result is less than the decision threshold, an assessment result of failing the evaluation is generated, indicating the existence of risk.

[0071] Typically, decision-making processes frequently involve scenarios requiring data processing and calculation (deriving indicators). The data processing logic is strongly correlated with business logic, and the task calculation logic differs across different decision flows. Furthermore, the actual operators of the system are often customer risk control personnel, who possess strong business knowledge but relatively weak IT skills. The introduction of a formula editor allows business personnel to configure the calculation logic of task nodes through a visual interface using specific syntax. However, editing formulas is still relatively cumbersome, involving input parameters, data types, and other data. Experience shows that formula formats are relatively fixed, with changes primarily occurring in the input parameters. Therefore, applying the decision model to the engine system's formula editor can increase the editor's flexibility and improve work efficiency.

[0072] In the technical solution of the decision model generation method provided in this embodiment of the invention, the meta-interpretive learning framework is trained by pre-acquired data samples and background knowledge to obtain a decision model generation model; the decision model generation model is generated based on the acquired operation and data type, which can increase the flexibility of the formula editor, improve the user's office efficiency, and enhance the user experience.

[0073] Figure 3 This is a schematic diagram of a decision model generation device provided in an embodiment of the present invention. This device is used to execute the aforementioned decision model generation method, such as... Figure 3 As shown, the device includes a training unit 11 and a generation unit 12.

[0074] Training unit 11 is used to train the meta-interpretive learning framework using pre-acquired data samples and background knowledge to obtain a decision model generation model.

[0075] The generation unit 12 is used to generate a model through the decision model, and generates a decision model based on the acquired operation and data type.

[0076] In this embodiment of the invention, the device further includes an acquisition unit 13 and a construction unit 14.

[0077] The acquisition unit 13 is used to acquire a specified number of data samples, which include input samples and expected formulas. The input samples include at least one operation and a data type.

[0078] The construction unit 14 is used to construct a feature information pool by using a specified logic programming language and based on preset function relationships.

[0079] In this embodiment of the invention, the training unit 11 is specifically used to train the meta-interpretive learning framework by inputting at least one operation, expectation formula, meta-rule and feature information pool into the meta-interpretive learning framework, and output the decision model generation model.

[0080] In this embodiment of the invention, the generation unit 12 is specifically used to input the operation and data type into the decision model to generate the model and output the decision model.

[0081] In this embodiment of the invention, the device further includes a calculation unit 15 and an evaluation unit 16.

[0082] The calculation unit 15 is used to input user data into the decision model for calculation and obtain decision results.

[0083] The evaluation unit 16 is used to evaluate the decision results according to the set evaluation rules and obtain the evaluation results.

[0084] In the embodiment of the present invention, the meta-interpretive learning framework is trained using pre-acquired data samples and background knowledge to obtain a decision model generation model; the decision model generation model generates a decision model based on the acquired operations and data types, which can increase the flexibility of the formula editor, improve the user's work efficiency, and enhance the user experience.

[0085] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer device, specifically, a computer device can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0086] This invention provides a computer device including a memory and a processor. The memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions. When the program instructions are loaded and executed by the processor, they implement the steps of the above-described decision model generation method. For a detailed description, please refer to the embodiments of the above-described decision model generation method.

[0087] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer device 600 suitable for implementing the embodiments of this application.

[0088] like Figure 4 As shown, the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate tasks and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the computer device 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0089] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal feedback (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed in storage section 608 as needed.

[0090] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611.

[0091] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0092] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0093] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0095] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0096] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0097] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0099] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0100] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A method for generating a decision model, characterized in that, The method includes: Get a specified number of data samples. The data samples include input samples and expected formulas. The input samples include operations and data types. The operations include greater than, sum, maximum value, average value and rounding. The data types include decimal and integer. The expected formula is the formula that the user expects to output. By specifying a logic programming language, a feature information pool is constructed based on preset operations; background knowledge includes the feature information pool and the selected meta-rules. The computational operations, expectation formulas, meta-rules, and feature information pools are input into the meta-interpretive learning framework for training, and the output decision model is generated as a model. The decision model is generated based on the acquired operations and data types. The decision model is then applied to the formula editor in the engine system of the anti-fraud field. User data is input into the decision model for calculation to obtain the decision result; The decision-making results are evaluated by setting evaluation rules, and the evaluation results are the anti-fraud evaluation results. The meta-rules include meta-rule headers and corresponding meta-rule bodies. The meta-interpretation learning framework is trained by inputting computational operations, expectation formulas, meta-rules, and a feature information pool, and outputs a decision model to generate a model. This process includes: using computational operations and their corresponding expectation formulas as atomic targets; matching the atomic targets with the meta-rule headers to obtain the corresponding meta-rule headers; matching the meta-rule bodies corresponding to the meta-rule headers with restricted clauses in the feature information pool; if the meta-rule body matches a restricted clause, saving the restricted clause; if the meta-rule body cannot match a restricted clause, using the meta-rule body as a new atomic target, and repeating the steps of matching the atomic targets with the meta-rule headers to obtain the corresponding meta-rule headers until matching is complete.

2. The decision model generation method according to claim 1, characterized in that, The step of generating a decision model through the decision model, based on the acquired computational operations and data types, includes: Input the computational operations and data types into the decision model to generate the model, and output the decision model.

3. The decision model generation method according to claim 1, characterized in that, Also includes: If the decision model generation model is not output within the set time limit, the data samples and background knowledge are adjusted, and the step of training the meta-interpretive learning framework with the pre-acquired data samples and background knowledge is re-executed to obtain the decision model generation model.

4. The decision model generation method according to claim 1, characterized in that, The number of input samples is 5 to 10.

5. A decision model generation device, characterized in that, The device includes: The acquisition unit is used to acquire a specified number of data samples. The data samples include input samples and expected formulas. The input samples include operation operations and data types. The operation operations include greater than, sum, maximum value, average value and rounding. The data types include decimal and integer. The expected formula is the formula that the user expects to output. The building unit is used to construct a feature information pool by using a specified logic programming language and based on preset operations; the background knowledge includes the feature information pool and the selected meta-rules. The training unit is used to input the computational operations, expectation formulas, meta-rules, and feature information pools into the meta-interpretive learning framework for training, and output the decision model generation model. The generation unit is used to generate a model through the decision model. Based on the acquired operation and data type, the decision model is generated and applied to the formula editor in the engine system of the anti-fraud field. The computing unit is used to input user data into the decision model for calculation and obtain decision results; The evaluation unit is used to evaluate the decision results according to the set evaluation rules, and obtain the evaluation results, which are the anti-fraud evaluation results. The meta-rule includes a meta-rule header and a corresponding meta-rule body. The training unit is specifically used to: take the operation and the corresponding expected formula as atomic targets; match the atomic targets with the meta-rule headers to obtain the corresponding meta-rule headers; match the meta-rule body corresponding to the meta-rule header with the restricted clauses in the feature information pool; if the meta-rule body matches a restricted clause, save the restricted clause; if the meta-rule body cannot match a restricted clause, take the meta-rule body as a new atomic target, and repeat the steps of matching the atomic target with the meta-rule header to obtain the corresponding meta-rule header until the matching is completed.

6. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the decision model generation method as described in any one of claims 1 to 4.

7. A computer device comprising a memory and a processor, the memory for storing information including program instructions, and the processor for controlling the execution of the program instructions, characterized in that, When the program instructions are loaded and executed by the processor, the decision model generation method according to any one of claims 1 to 4 is implemented.

Images