An iterative optimization method and system for anti-risk control model parameters of accounting cases

Abstract

The present invention provides an accounting case wind control model parameter iterative optimization method and system, including: Step S1: Set parameter class special identifiers in the code statistical screening condition line; Step S2: The parameter class special identifiers in the code statistical screening condition line trigger special identifier dynamic parameter sensors; Step S3: Add the function of automatically generating explanatory text documents; Step S4: Set a visual configuration function area to configure the threshold of dynamic parameters; Step S5: Build the collaborative function of the job batch processing monitoring platform and the model development platform. The present invention has high accuracy, reduces the processing difficulty and error rate and saves time; has strong reusability, and both developers and testers can dynamically parameterize similar statistical screening condition dynamic parameters in the rules to achieve consistent modification of the same parameter, reducing the workload of code optimization.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a method and system for iteratively optimizing the parameters of an accounting case anti-risk control model. Background Art

[0002] The accounting case anti-risk data analysis system mainly undertakes the model construction and risk verification work in the field of accounting internal control of our bank. In the original internal control rule development process, business experts put forward refined requirements for internal control scenarios, and developers completed the development of requirements based on computer languages and scripts and cooperated with business testers (hereinafter referred to as testers) to complete the accuracy testing and optimization of the model. With the call for refined and diversified internal control scenarios and the need for continuous iterative upgrading of prevention and control standards, it adds many obstacles to the smooth progress and iterative optimization of the accounting case anti-risk project. From the perspectives of business and technology, they are respectively reflected in:

[0003] For business personnel, in the development process of expert rules, variable parameters need to be continuously adjusted to find the optimal value to make the model achieve the best performance. The optimization process takes a long time and has many iterations, and the parameter changes involved are frequent, which brings many troubles to the business testers. On the one hand, the business testers are not as familiar with the code as the developers, and they cannot quickly locate or even misjudge the target statistical screening conditions in the joint debugging test, resulting in a sharp increase in the testing difficulty; on the other hand, the thresholds of the parameters that need to be adjusted in the expert rules have strong reusability, and testers often need to make repetitive modifications to the parameters with the same assignment but different storage locations in the code, which greatly increases the testing time and the probability of making mistakes.

[0004] For developers, on the one hand, after the business testers complete the testing and feedback to the business experts, corresponding modifications need to be made before going online according to the new parameter thresholds given by the business experts, and the iteration of the optimization content needs to be achieved by modifying the original code, which has defects such as low efficiency and weak reusability; on the other hand, the code of the expert rules released after going online cannot be changed. When it is necessary to optimize the rules due to the increase in batch data and the continuous enrichment of actual internal control scenarios, the code must be refactored to build a new model to complete the iterative upgrade of the internal control rules, which will increase the time cost by 2 to 4 weeks in the actual process and the efficiency is too low.

[0005] Therefore, there is an urgent need for a new technical solution to reduce the technical threshold of the business side, improve the joint debugging efficiency of business testers, and at the same time assist developers in low-code development, improve the post-go-live rule adjustment mechanism, and enhance the development efficiency. Summary of the Invention

[0006] Aiming at the defects in the prior art, the purpose of the present invention is to provide a method and system for iteratively optimizing the parameters of an accounting case anti-risk control model.

[0007] An iterative optimization method for anti-risk control model parameters provided by the present invention includes:

[0008] Step S1: Set a parameter class special identifier in the code statistics screening condition line;

[0009] Step S2: The special identifier of the parameter class in the code statistics screening condition line triggers a special identifier dynamic parameter sensor;

[0010] Step S3: Add a function for automatically generating explanatory text documents;

[0011] Step S4: Set a visual configuration function area to configure the threshold of dynamic parameters;

[0012] Step S5: Build a collaborative function between the job batch processing monitoring platform and the model development platform.

[0013] Preferably, in the step S1:

[0014] The standard input format is set as the left half identifier + dynamic parameter index number + colon + initial parameter threshold + the right half identifier.

[0015] Preferably, in the step S2:

[0016] When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line and automatically record it in the text document, and select whether to enable this function in the UI interface;

[0017] When the special identifier dynamic parameter sensor is triggered, record the parameter index number in the identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0018] Preferably, in the step S3:

[0019] When the special identifier dynamic parameter sensor is triggered, the function for automatically generating the description document is triggered, and the English name of the parameter and the note of this line are sequentially recorded in the text document;

[0020] When the tester accurately locates the statistical screening conditions and the corresponding parameters, the description document function can be closed.

[0021] Preferably, in the step S5:

[0022] Build a collaborative function between the job batch processing monitoring platform and the model development platform by writing shell scripts and computer language function scheduling interfaces;

[0023] After the risk expert rules on the model development platform are launched, by passing parameters through the script of the batch monitoring platform, the limitation of the lock protection added after the expert rule code is launched in the model development platform is bypassed, and the optimization iteration of the expert rule is realized from the side.

[0024] The standard input format of the script in the batch monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer;

[0025] The script written in the script interface module of the batch scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

[0026] According to an accounting case anti-risk control model parameter iteration optimization system provided by the present invention, it includes:

[0027] Module M1: Set a parameter class special identifier in the code statistics screening condition line;

[0028] Module M2: The special identifier of the parameter class in the code statistics screening condition line triggers a special identifier dynamic parameter sensor;

[0029] Module M3: Add a function of automatically generating an explanatory text document;

[0030] Module M4: Set a visual configuration function area to configure the threshold of dynamic parameters;

[0031] Module M5: Build a collaborative function between the job batch monitoring platform and the model development platform.

[0032] Preferably, in the module M1:

[0033] The standard input format is set as the left half identifier + dynamic parameter index number + colon + initial parameter threshold + right half identifier.

[0034] Preferably, in the module M2:

[0035] When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line and automatically record it in the text document, and select whether to enable this function in the UI interface;

[0036] When the special identifier dynamic parameter sensor is triggered, record the parameter index number in the identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0037] Preferably, in the module M3:

[0038] When the special identifier dynamic parameter sensor is triggered, the function of automatically generating the instruction document is triggered, and the English name of the parameter and the remarks of this line are recorded in the text document in sequence;

[0039] When the tester accurately locates the statistical screening conditions and corresponding parameters, the function of the instruction document can be closed.

[0040] Preferably, in the module M5:

[0041] By writing shell scripts and computer language function scheduling interfaces, the collaborative function of the job batch processing monitoring platform and the model development platform is constructed;

[0042] After the risk expert rules on the model development platform are launched, by means of passing parameters to the batch processing monitoring platform script, the limitation of the lock protection after the expert rule code is launched on the model development platform is bypassed, and thus the optimization and iteration of the expert rules are realized from the side;

[0043] The standard input format of the script in the batch processing monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer;

[0044] The script written in the script interface module of the batch processing scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention has high accuracy. The business tester can accurately and quickly locate the statistical screening conditions in the expert rules through the instruction document generated by the system, reducing the processing difficulty and error rate and saving time;

[0047] 2. The present invention has strong reusability. Both developers and testers can dynamically parameterize the similar statistical screening conditions in the rules to achieve the consistent modification of the same parameter, reducing the workload of code optimization;

[0048] 3. The present invention has high convenience. Both developers and testers can configure the parameters in the visual configuration function area, simplifying the existing parameter configuration process and improving the rule optimization efficiency;

[0049] 4. The present invention has strong scalability. After the expert rules are launched, developers can still use the batch processing script scheduling method. On the one hand, they can iteratively optimize the statistical parameters, which plays an important guiding role in the security and stable operation of the data; on the other hand, they can also expand the time parameters to realize functions such as filling in missing numbers;

[0050] 5. The present invention can improve the testing efficiency of business testers and reduce the error rate; improve the optimization efficiency of developers; have the ability to expand the platform, bypass the locking limitation after the rules are launched, improve the optimization and iteration efficiency, and ultimately achieve the effect of reducing the technical barriers of business testers, realizing low-code and high-efficiency in the optimization process, thereby achieving efficiency increase and cost reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Other features, objects, and advantages of the present invention will become more apparent by reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0052] Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The present invention will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those of ordinary skill in the art can make several changes and improvements without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0054] Embodiment 1:

[0055] The purpose of the present invention is to provide a method and system for iterative optimization of accounting case anti-risk control model parameters. By writing computer functions and interfaces, a plurality of modules with synergistic effects are designed to construct a parameter iterative optimization system.

[0056] Specifically: 1 - The parameter identification module, the automated specification document generation module, and the visual configuration parameter passing interface work together to achieve the effects of accurately positioning the parameter position, low-code development, and unified parameter passing;

[0057] 2 - The parameter passing interface module and the batch processing scheduling script interface module cooperate with each other to achieve the effect of being able to optimize the rules even after the risk expert rules are launched;

[0058] The development of accounting case anti-internal control rules is on a multi-dimensional data analysis platform. The code model development is encapsulated in the canvas. The canvas can only be viewed and cannot be edited after going online (due to the system design for security reasons, so the code cannot be directly modified after going online); the role of the batch processing scheduling platform is that the batch processing platform automatically processes the pushed data through the code model every day. After adding a new script interface module to the batch processing platform, it is now possible to directly pass parameters from the batch processing platform to the parameter passing interface in the visual function configuration area of the canvas.

[0059] According to an accounting case anti-risk control model parameter iterative optimization method provided by the present invention, as Figure 1 shown, it includes:

[0060] Step S1: Set the parameter class special identifier in the code statistics and screening condition line;

[0061] Specifically, in the step S1:

[0062] The standard input format is set as the left - hand identifier + dynamic parameter index number + colon + initial parameter threshold + right - hand identifier.

[0063] Step S2: The parameter class special identifier in the code statistics and screening condition line triggers the special identifier dynamic parameter sensor;

[0064] Specifically, in the step S2:

[0065] When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line, automatically record it in the text document, and select whether to enable this function in the UI interface;

[0066] When the special identifier dynamic parameter sensor is triggered, record the parameter index number in this identifier. The addresses pointed to by the same parameter index number are the same, and generate an interface to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0067] Step S3: Add the function of automatically generating an explanatory text document;

[0068] Specifically, in the step S3:

[0069] When the special identifier dynamic parameter sensor is triggered, the function of automatically generating an instruction document is triggered, and record the English name of the parameter and the note of this line in the text document in sequence;

[0070] When the tester accurately locates the statistical screening conditions and corresponding parameters, the function of the instruction document can be turned off.

[0071] Step S4: Set the visual configuration function area to configure the threshold of the dynamic parameter;

[0072] Step S5: Build the collaborative function between the job batch processing monitoring platform and the model development platform.

[0073] Specifically, in the step S5:

[0074] Build the collaborative function between the job batch processing monitoring platform and the model development platform by writing shell scripts and computer language function scheduling interfaces;

[0075] After the risk expert rules on the model development platform are launched, by passing parameters through the script of the batch monitoring platform, the limitation of the lock protection added after the expert rule code is launched in the model development platform is bypassed, and the optimization iteration of the expert rule is realized from the side.

[0076] The standard input format of the script in the batch monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer;

[0077] The script written in the script interface module of the batch scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

[0078] Embodiment 2:

[0079] Embodiment 2 is a preferred example of Embodiment 1 to illustrate the present invention more specifically.

[0080] The present invention also provides an accounting case anti-risk control model parameter iteration optimization system, and the accounting case anti-risk control model parameter iteration optimization system can be realized by executing the process steps of the accounting case anti-risk control model parameter iteration optimization method, that is, those skilled in the art can understand the accounting case anti-risk control model parameter iteration optimization method as a preferred implementation manner of the accounting case anti-risk control model parameter iteration optimization system.

[0081] According to an accounting case anti-risk control model parameter iteration optimization system provided by the present invention, it includes:

[0082] Module M1: Set a parameter class special identifier in the code statistics and screening condition line;

[0083] Specifically, in the module M1:

[0084] The standard input format is set as the left half identifier + dynamic parameter index number + colon + initial parameter threshold + right half identifier.

[0085] Module M2: The parameter class special identifier in the code statistics and screening condition line triggers a special identifier dynamic parameter sensor;

[0086] Specifically, in the module M2:

[0087] When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line, automatically record it in the text document, and select whether to enable this function in the UI interface;

[0088] When the special identifier dynamic parameter sensor is triggered, record the parameter index number in the identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0089] Module M3: Add the function of automatically generating explanatory text documents;

[0090] Specifically, in the said Module M3:

[0091] When the special identifier dynamic parameter sensor is triggered, the function of automatically generating the explanatory document is triggered, and the English name of the parameter and the note of this line are recorded in sequence in the text document;

[0092] When the tester accurately locates the statistical screening conditions and corresponding parameters, the function of the explanatory document can be turned off.

[0093] Module M4: Set a visual configuration function area to configure the threshold of dynamic parameters;

[0094] Module M5: Build the collaborative function of the job batch processing monitoring platform and the model development platform.

[0095] Specifically, in the said Module M5:

[0096] Build the collaborative function of the job batch processing monitoring platform and the model development platform by writing shell scripts and computer language function scheduling interfaces;

[0097] After the risk expert rules on the model development platform are launched, by means of passing parameters to the batch processing monitoring platform script, bypass the limitation of the lock protection after the expert rule code on the model development platform is launched, and then realize the optimization and iteration of the expert rules from the side;

[0098] The standard input format of the script in the batch processing monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer;

[0099] The script written in the script interface module of the batch processing scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

[0100] Embodiment 3:

[0101] Embodiment 3 is a preferred example of Embodiment 1 to more specifically illustrate the present invention.

[0102] Aiming at the defects in the prior art, the purpose of the present invention is to provide an accounting case anti-risk control model parameter iteration optimization method and system. By expanding the functions of the multi-dimensional data analysis platform and the batch processing scheduling platform, it optimizes the original development process and improves the accounting case anti-internal control rule development system, including:

[0103] Step S1: During the development process, the developer adds parameter class identifiers to the code;

[0104] Step S2: Add the function of identifying specific dynamic parameter identifiers by writing computer language functions;

[0105] Step S3: Add the function of automatically generating explanatory text documents by writing computer language functions.

[0106] Specifically, when S2 is triggered, the code comment part of the corresponding line is marked and automatically recorded in the text document. To avoid wasting resources, you can choose whether to enable this function in the UI interface;

[0107] Specifically, when S2 is triggered, record the parameter index number in this identifier. The addresses pointed to by the same parameter index number are the same, and an interface to the visual configuration function area is generated, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0108] Step S4: Set a visual configuration function area in the code development interface, and configure the threshold of the dynamic parameter in the visual configuration area.

[0109] Step S5: Build the collaborative function of the job batch processing monitoring platform and the model development platform by writing shell scripts and computer language function scheduling interfaces.

[0110] Specifically, after the risk expert rules on the model development platform are launched, by passing parameters to the batch processing monitoring platform script, the limitation of the lock protection after the expert rule code is launched on the model development platform can be bypassed, and thus the optimization and iteration of the expert rule can be achieved from the side.

[0111] Specifically, the statistical screening conditions are based on the actual combat experience of business experts.

[0112] Specifically, in step S1, the standard input format is set as the left half identifier (default <<<) + dynamic parameter index number (default capital English characters, such as A, B, C...) + colon + initial parameter threshold + right half identifier (default >>>).

[0113] Specifically, in step S3, when the tester accurately locates the statistical screening conditions and corresponding parameters, the description document function can be turned off to save system resources.

[0114] Specifically, in step S5, the standard input format of the script in the batch processing monitoring platform should be the job production cluster server address + production job number + production job date + dynamic parameter transfer.

[0115] In this optimized system, when the developer completes the code and runs the code according to the regulations:

[0116] The identifier in step 1 is triggered by the recognition function in step 2;

[0117] When step 1 is triggered, the function for automatically generating a description document in step 3 is triggered, and the English name of the parameter and the note for that line are recorded in sequence in a text document (for example: "Parameter A Today_avg_pay The payment amount per credit transaction of the account today is greater than 5000 yuan");

[0118] When step 1 is triggered, the parameters in the visual configuration area in step 4 are mapped one-to-one with the parameter values in the identifier in step 1 in sequence. Passing the parameter values through the interface in the visual configuration area can simplify the development process (for example: continuing from the previous example, when we write the value 10000 to parameter A in the visual configuration area and run the code, at this time, 5000 of parameter A is replaced by 10000);

[0119] In step 5, the script written in the script interface module of the batch processing scheduling platform assigns the parameter values to the parameters in the visual function configuration area in step 4.

[0120] Embodiment 4:

[0121] Embodiment 4 is a preferred example of Embodiment 1 to more specifically illustrate the present invention.

[0122] An optimization iteration method for accounting case prevention data analysis risk rule parameters, including:

[0123] Step S1: Developers set parameter class special identifiers in the code statistical screening condition line according to the specified standard format and complete the code development; the statistical screening condition line is used to describe the degree of risk confidence intervals existing in each risk scenario, and the special identifier is an information symbol for parameter identification and indexing. For example: the statistical screening condition line "The payment amount per credit transaction of the account today is greater than <<<A: 5000 yuan>>>", it should be understood that under the design of business experts, when "the payment amount per credit transaction of a specific account" today is greater than "the payment amount per credit transaction of an account in the normal scenario", there is a risk problem for this account. Among them, "<<<>>>" is the special identifier, and A in the special identifier represents parameter A (that is, "the payment amount per credit transaction of an account in the normal scenario"), and 5000 yuan is the threshold set for this parameter (based on expert experience).

[0124] Step S2: The tester enables the function of automatically generating a description document (step S3 of the invention content), runs the code block, and obtains the description document.

[0125] Step S3: The tester conducts parameter passing and testing in the visual configuration function area, communicates with the business expert, and feeds back the final modification opinions to the developer (step S4 of the invention content)

[0126] Step S4: Developers modify the code according to the final feedback from the business side using the visual configuration ribbon and complete the final version of the code.

[0127] Step S5: Developers deploy the expert rules to the production environment. According to subsequent feedback, in the shell script scheduling module of the batch processing monitoring and scheduling platform, according to the specified input format: job production cluster server address - production job number (e.g., 3331) - production job date (e.g., 20220915) - dynamic parameter passing (e.g., passing in 2 types of dynamic parameters, namely the ratio of debit amount to total amount 0.7 and the ratio of credit amount to total amount 0.5, and the format is corresponding to: 0.7,0.5), cross-platform rule optimization is achieved.

[0128] Embodiment 5:

[0129] Embodiment 5 is a preferred example of Embodiment 1 to more specifically illustrate the present invention.

[0130] Aiming at the defects in the prior art, the object of the present invention is to provide an optimization iteration method and system for accounting case prevention data analysis risk rule parameters

[0131] According to an optimization iteration method for accounting case prevention data analysis risk rule parameters provided by the present invention, it includes:

[0132] Step S1: Set parameter class special identifiers in the code statistical screening condition line according to the specified standard format;

[0133] Step S2: By writing a computer language function, add the function of a special identifier dynamic parameter sensor, and the parameter class special identifier in the code statistical screening condition line triggers the sensor;

[0134] Step S3: By writing a computer language function, add the function of automatically generating an explanatory text document.

[0135] Specifically, when the dynamic parameter sensor in S2 is triggered, intercept the statistic in the statistical screening condition line where the S2 sensor is triggered and feedback the statistic description of the code comment in this line, and automatically record it in the text document. To avoid resource waste, it is possible to select whether to enable this function at the UI interface;

[0136] Step S4: Set a visual configuration ribbon in the code development interface, and configure the threshold of the dynamic parameter in the visual configuration area.

[0137] Specifically, through the indicative trigger of the sensor in S2, record the parameter index number in the identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration ribbon, so that when passing parameters in the visual configuration area, all parameters with the same index number can achieve the purpose of unified modification.

[0138] Step S5: Build the collaborative function between the job batch monitoring platform and the code development platform by writing shell scripts and computer language function scheduling interfaces.

[0139] Specifically, after the risk expert rules on the code development platform are launched, the limitation of the lock protection added after the launch of the expert rule code in the code development platform can be bypassed by passing parameters to the batch monitoring platform script, thereby realizing the optimization and iteration of the expert rules from the side.

[0140] An accounting case prevention data analysis risk rule parameter optimization and iteration system provided by the present invention includes:

[0141] Module M1: Set parameter class special identifiers in the code statistics screening condition line according to the specified standard format;

[0142] Module M2: Add the function of a special identifier dynamic parameter sensor by writing computer scripts, and trigger the sensor with the parameter class special identifier in the code statistics screening condition line;

[0143] Module M3: Add the function of automatically generating an explanatory text document by writing computer scripts. Specifically, when the dynamic parameter sensor in S2 is triggered, intercept the statistic in the statistic screening condition line triggered by the S2 sensor and feedback the statistic description of the code comment of this line, and automatically record it in the text document. To avoid resource waste, it is possible to select whether to enable this function at the UI interface;

[0144] Module M4: Set a visual configuration function area in the code development interface, and configure the threshold of the dynamic parameter in the visual configuration area. Specifically, through the indicative trigger of the sensor in S2, record the parameter index number in the identifier, the addresses pointed to by the same parameter index number are the same, and generate an interface to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified.

[0145] Module M5: Build the collaborative function of batch script and statistical screening condition dynamic parameterization scheduling. After the risk expert rules are launched, the limitation of the lock added after the launch of the expert rule code can still be bypassed by passing parameters to the script, so as to realize the optimization and iteration of the expert rules.

[0146] Those skilled in the art know that, in addition to implementing the systems, devices and their respective modules provided by the present invention in the form of pure computer-readable program codes, the method steps can be logically programmed to enable the systems, devices and their respective modules provided by the present invention to be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Therefore, the systems, devices and their respective modules provided by the present invention can be regarded as a kind of hardware components, and the modules included therein for implementing various programs can also be regarded as the structures within the hardware components; the modules for implementing various functions can also be regarded as either software programs for implementing the methods or the structures within the hardware components.

[0147] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. An iterative optimization method for accounting case anti-risk control model parameters, characterized in that Including: Step S1: Set a parameter class special identifier in the code statistics screening condition line; Step S2: The special identifier in the code statistics screening condition line triggers a special identifier dynamic parameter sensor; Step S3: Add a function for automatically generating an explanatory text document; Step S4: Set a visual configuration function area to configure the threshold of dynamic parameters; Step S5: Build the collaborative function of the job batch processing monitoring platform and the model development platform; In the said Step S2: When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line, automatically record it in the text document, and select whether to enable this function in the UI interface; When the special identifier dynamic parameter sensor is triggered, record the parameter index number in this identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified; In the said Step S5: Build the collaborative function of the job batch processing monitoring platform and the model development platform by writing shell scripts and computer language function scheduling interfaces; After the risk expert rules on the model development platform are launched, by passing parameters to the batch processing monitoring platform script, bypass the limitation of the lock protection after the expert rule code is launched on the model development platform, and then realize the optimization and iteration of the expert rule from the side.

2. The accounting case wind control model parameter iteration optimization method according to claim 1, characterized in that In the said Step S1: The standard input format is set as the left half identifier + dynamic parameter index number + colon + initial parameter threshold + right half identifier.

3. An accounting case wind control model parameter iterative optimization method according to claim 1, characterized in that, In the said Step S3: When the special identifier dynamic parameter sensor is triggered, the function for automatically generating an instruction document is triggered, and the English name of the parameter and the note of this line are recorded in the text document in sequence; When the tester accurately locates the statistical screening conditions and corresponding parameters, the instruction document function can be closed.

4. The accounting case anti-risk control model parameter iterative optimization method according to claim 1, characterized in that, In the said Step S5: The standard input format of the script in the batch processing monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer; The script written in the script interface module of the batch processing scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

5. An accounting case wind control model parameter iterative optimization system, characterized in that, Including: Module M1: Set a parameter class special identifier in the code statistics screening condition line; Module M2: The special identifier in the code statistics screening condition line triggers a special identifier dynamic parameter sensor; Module M3: Add a function for automatically generating an explanatory text document; Module M4: Set a visual configuration function area to configure the threshold of dynamic parameters; Module M5: Build the collaborative function of the job batch processing monitoring platform and the model development platform; In the said Module M2: When the special identifier dynamic parameter sensor is triggered, mark the code comment part of the corresponding line, automatically record it in the text document, and select whether to enable this function in the UI interface; When the special identifier dynamic parameter sensor is triggered, record the parameter index number in this identifier. The addresses pointed to by the same parameter index number are the same, and an interface is generated to the visual configuration function area, so that when passing parameters in the visual configuration area, all parameters with the same index number can be uniformly modified; In the module M5: By writing shell scripts and computer language function scheduling interfaces, the collaborative function between the job batch monitoring platform and the model development platform is constructed; After the risk expert rules on the model development platform are launched, by passing parameters to the script of the batch monitoring platform, the limitation of the lock protection after the expert rule code is launched on the model development platform is bypassed, and thus the optimization and iteration of the expert rules are realized from the side.

6. The anti-risk control model parameter iterative optimization system for accounting cases according to claim 5, characterized in that, In the module M1: The standard input format is set as the left - hand identifier + dynamic parameter index number + colon + initial parameter threshold + right - hand identifier.

7. The accounting case anti-risk control model parameter iterative optimization system according to claim 5, characterized in that In the module M3: When the special identifier dynamic parameter sensor is triggered, the function of automatically generating the instruction document is triggered, and the English name of the parameter and the note of that line are recorded in the text document in sequence; When the tester accurately locates the statistical screening conditions and corresponding parameters, the function of the instruction document can be turned off.

8. An accounting case wind control model parameter iterative optimization system according to claim 5, characterized in that In the module M5: The standard input format of the script in the batch monitoring platform is: job production cluster server address + production job number + production job date + dynamic parameter transfer; The script written in the script interface module of the batch scheduling platform will assign the parameter value to the parameter in the visual function configuration area.

Images