Conformity training and conformity judgment method, device, equipment and medium

Abstract

The application discloses a compliance training and judgment method and device, equipment and medium. The application relates to the technical field of big data. The method comprises the following steps: acquiring a compliance requirement file, historical business data and a compliance result label of the historical business data; determining a to-be-audited training sample according to the compliance requirement file and the historical business data; and training a pre-constructed compliance model according to the to-be-audited training sample and the corresponding compliance result label. According to the technical scheme, the trained compliance model has higher accuracy when performing compliance review, can realize compliance review of multiple compliance requirement files, has a wider application range, does not need to extract and construct multiple auditing rules, and is more convenient to use.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of big data, and in particular to a compliance model training and compliance judgment method and device, equipment and a medium. BACKGROUND

[0002] In the process of applying for a service, it is often necessary to conduct compliance review on the applied service according to various compliance requirement files.

[0003] The existing compliance review mainly adopts a manual review mode. In the review process, the data to be reviewed and the compliance requirement files to be referred to are complicated, and the review rules in the compliance requirement files are various, resulting in low review efficiency. At the same time, the manual review mode has strong subjectivity and low accuracy of the review results. SUMMARY

[0004] Embodiments of the present application provide a compliance model training and compliance judgment method, device, equipment and medium to improve the accuracy and review efficiency of compliance review.

[0005] In a first aspect, embodiments of the present application provide a compliance model training method, which comprises:

[0006] obtaining a compliance requirement file, historical business data and a compliance result label of the historical business data;

[0007] determining a training sample to be reviewed according to the compliance requirement file and the historical business data;

[0008] training a pre-constructed compliance model according to the training sample to be reviewed and the corresponding compliance result label.

[0009] In a second aspect, embodiments of the present application provide a compliance judgment method, which comprises:

[0010] determining business data to be reviewed of a business to be reviewed and a compliance requirement file to be reviewed of the business data to be reviewed;

[0011] determining a prediction sample to be reviewed according to the compliance requirement file to be reviewed and the business data to be reviewed;

[0012] inputting the prediction sample to be reviewed into the trained compliance model to obtain a compliance result; wherein the compliance model is trained based on the method of the first aspect.

[0013] In a third aspect, embodiments of the present application further provide a compliance model training device, which comprises:

[0014] a data acquisition module configured to obtain a compliance requirement file, historical business data and a compliance result label of the historical business data;

[0015] The training sample determination module is configured to determine the training sample to be audited according to the compliance requirement file and the historical business data.

[0016] The training module is configured to train the pre-constructed compliance model according to the training sample to be audited and the corresponding compliance result label.

[0017] In a fourth aspect, the embodiments of the present application further provide a compliance judgment device, which comprises:

[0018] The compliance requirement file determination module is configured to determine the business data to be audited of the business to be audited and the compliance requirement file to be audited of the business data to be audited.

[0019] The prediction sample determination module is configured to determine the prediction sample to be audited according to the compliance requirement file to be audited and the business data to be audited.

[0020] The compliance result determination module is configured to input the prediction sample to be audited into the trained compliance model to obtain a compliance result, wherein the compliance model is trained based on the device of the third aspect.

[0021] In a fifth aspect, the embodiments of the present application further provide an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the compliance model training method or the compliance judgment method of any of the embodiments of the present application when executing the program.

[0022] In a sixth aspect, the embodiments of the present application further provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the compliance model training method or the compliance judgment method of any of the embodiments of the present application.

[0023] In a seventh aspect, the embodiments of the present application further provide a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the compliance model training method or the compliance judgment method of any of the embodiments of the present application.

[0024] The technical scheme of the embodiments of the present application determines the training sample to be audited according to the compliance requirement file and the historical business data, trains the pre-constructed compliance model according to the compliance result label of the training sample to be audited and the historical business data, so that the trained compliance model has the compliance review capability of multiple compliance requirement files, thereby achieving higher accuracy when the compliance model is used for subsequent compliance review, and achieving compliance review of multiple compliance requirement files, and having a wider application range. In addition, when performing compliance review, there is no need to summarize and refine the audited rules, so that the compliance review process is more convenient. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0026] Figure 1 A flowchart of a compliance training method provided by an embodiment of the present application is shown in FIG. 1.

[0027] Figure 2 A flowchart of another compliance training method provided by an embodiment of the present application is shown in FIG. 2.

[0028] Figure 3 A flowchart of still another compliance training method provided by an embodiment of the present application is shown in FIG. 3.

[0029] Figure 4 A flowchart of a compliance judgment method provided by an embodiment of the present application is shown in FIG. 4.

[0030] Figure 5 A flowchart of another compliance judgment method provided by an embodiment of the present application is shown in FIG. 5.

[0031] Figure 6 A structural diagram of a compliance training device provided by an embodiment of the present application is shown in FIG. 6.

[0032] Figure 7 A structural diagram of a compliance judgment device provided by an embodiment of the present application is shown in FIG. 7.

[0033] Figure 8 A structural diagram of an electronic device provided by an embodiment of the present application is shown in FIG. 8. DETAILED DESCRIPTION

[0034] The present application will be further described in detail below in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and not to limit the present application. In addition, it should be noted that only the parts related to the present application are shown in the drawings, and not all the structures.

[0035] It should be noted that similar reference numerals and letters refer to like items in the accompanying drawings, and once an item is defined in one drawing, it is not necessary to further define and explain it in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", and the like are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance. The acquisition, storage, use, processing, etc. of compliance requirement files, historical business data, compliance result labels, to-be-audited compliance requirement files, and to-be-audited business data in the technical solutions of the present application all comply with the relevant provisions of national laws and regulations.

[0036] The embodiments of the present application are applicable to the scenario of compliance review of business data to improve the efficiency and accuracy of compliance review. In order to facilitate understanding, first, the training process of the compliance model involved in the embodiments of the present application is described in detail. The compliance model can be a neural network model. The compliance model is obtained by model training.

[0037] Figure 1 A flowchart of a compliance model training method provided by the embodiments of the present application, the embodiments of the present application are applicable to the case of training the compliance model for compliance review. The method can be executed by a compliance model training device, which can be realized by software and / or hardware, and can be configured in an electronic device.

[0038] As shown in the compliance model training method, the specific steps include the following steps: Figure 1

[0039] S110, obtaining compliance requirement files, historical business data, and compliance result labels of the historical business data.

[0040] The compliance requirement file refers to the restriction conditions that should be met for the corresponding business standardized by the organization or group with supervisory responsibilities in an authoritative form. The number of compliance requirement files is at least one. It should be noted that the corresponding compliance requirement files are different for different business scenarios. In a specific implementation manner, the compliance requirement file can be a file for limiting enterprise loans and the like.

[0041] It can be understood that the business data required for a business is specified in the compliance requirement file. The business data is used to represent the basic data required to be provided by the business initiator when initiating the related business. The historical business data is used to represent the business data of the business initiator whose business has been processed. The business data can include a business field and a field value corresponding to the business field. Specifically, in the loan business scenario, the business field can be at least one of loan amount, loan period, and loan interest rate.

[0042] ​The compliance result label is used to represent the compliance review result of the historical business data. Specifically, the compliance result label can be compliant or non-compliant. It can be understood that in the compliance requirement file, the compliance requirement for reviewing the business data is included. The compliance result label is compliant, indicating that the historical business data meets the compliance requirement for reviewing the business data in the compliance requirement file; and the compliance result label is non-compliant, indicating that the historical business data does not meet the compliance requirement for reviewing the business data in the compliance requirement file.

[0043] S120, determining the training sample to be audited according to the compliance requirement file and the historical business data.

[0044] The training sample to be audited is the input data when the compliance model is trained, and the number of the training sample to be audited is at least one, usually multiple. Specifically, according to the compliance requirement file and the historical business data, the associated data of the compliance requirement file in the historical business data can be determined, so as to determine the training sample to be audited according to the associated data.

[0045] S130, training the pre-constructed compliance model according to the training sample to be audited and the corresponding compliance result label.

[0046] The corresponding compliance result label refers to the compliance result label corresponding to the training sample to be audited. It can be understood that the training sample to be audited and the compliance result label of the training sample to be audited can be used as a training sample pair to train the pre-constructed compliance model to adjust the network parameters in the model. The pre-constructed compliance model refers to a compliance model that has not been trained, which can be a neural network model or other models, and the specific network structure of the model is not limited in the present application.

[0047] Specifically, the training sample to be audited is input into the pre-constructed compliance model to obtain a model prediction result, and the network parameters in the compliance model are adjusted according to the difference between the model prediction result and the compliance result label of the corresponding training sample to be audited until the training termination condition is met. The training termination condition can include at least one of the following: the number of training samples to be audited reaches a preset number threshold, the accuracy of the model reaches a preset accuracy threshold, and the compliance model tends to converge.

[0048] The technical scheme of the embodiment of the application determines the to-be-audited training sample according to the compliance requirement file and the historical business data, trains the pre-constructed compliance model according to the compliance result label of the to-be-audited training sample and the historical business data, so that the trained compliance model has the compliance review capability of multiple compliance requirement files, thereby the accuracy is higher when the compliance model is used for compliance review subsequently, and the compliance review of multiple compliance requirement files can be implemented, and the application range is wider. In addition, the summary and extraction of the audited rule are not needed when the compliance review is performed, so that the compliance review process is more convenient.

[0049] Optionally, the to-be-audited training sample is determined according to the compliance requirement file and the historical business data, including: obtaining the initiator category of the business initiator corresponding to the historical business data; and determining the to-be-audited training sample according to the initiator category corresponding to the historical business data, the compliance requirement file and the historical business data.

[0050] The business initiator is used for uniquely representing the subject of initiating the business. For example, the business initiator can be an enterprise, an organization or an individual. The historical business data corresponds to the business initiator, and at least one piece of historical business data can correspond to one business initiator; the historical business data corresponding to different business initiators can be the same or different. The initiator category is used for representing the category to which the business initiator belongs. The initiator category can be set according to actual conditions. For example, the initiator category can be a large customer, a small customer, a large enterprise or a small enterprise.

[0051] Specifically, the compliance requirement file corresponding to the initiator category is selected from a large number of compliance requirement files according to the initiator category of the business initiator corresponding to the historical business data, and the to-be-audited training sample is determined according to the associated data corresponding to the historical business data in the compliance requirement file corresponding to the initiator category.

[0052] By introducing the initiator category of the business initiator corresponding to the historical business data into the to-be-audited training sample when the model is trained, the trained model can have the distinguishing ability of the business initiator, so that the compliance review can be performed in a targeted manner for different categories of business initiators in the compliance review process, and the flexibility and accuracy of the compliance review are improved.

[0053] It should be noted that the compliance requirement file can be updated or adjusted according to actual conditions. Therefore, when the compliance model training is performed, the updated compliance requirement file, the historical business data and the compliance result label of the historical business data can also be obtained; the to-be-audited training sample is determined according to the updated compliance requirement file and the historical business data; and the trained compliance model is trained again according to the to-be-audited training sample and the corresponding compliance result label. In the above manner, the trained compliance model can continuously adapt to the new compliance requirement file, which helps to improve the adaptability of the compliance model to the new compliance requirement file and the accuracy of the compliance review result.

[0054] On the basis of each of the technical solutions described above, the present application further provides an optional embodiment. In the optional embodiment, the operation of “determining the to-be-audited training sample according to the compliance requirement file and the historical business data” is specifically embodied as “taking the historical business data corresponding to the compliance requirement file as a reference file, and determining a historical to-be-evaluated field in the reference file; determining a field value corresponding to the historical to-be-evaluated field in the historical business data; and generating the to-be-audited training sample according to the historical to-be-evaluated field and the field value of the historical to-be-evaluated field”, so as to enrich the determination mechanism of the to-be-audited training sample. It should be noted that the parts not described in detail in the embodiments of the present application can be referred to the related descriptions of other embodiments.

[0055] Referring to Figure 2 The compliance model training method shown in FIG. 2 includes the following specific steps:

[0056] S210, obtaining a compliance requirement file, historical business data and a compliance result label of the historical business data.

[0057] S220, taking the historical business data corresponding to the compliance requirement file as a reference file, and determining a historical to-be-evaluated field in the reference file.

[0058] The reference file is used to represent the compliance requirement file that needs to be used when the historical business data is subjected to compliance review. It can be understood that the historical to-be-evaluated field in the reference file is the requirement for the compliance review of the historical business data.

[0059] For example, the compliance requirement file corresponding to the historical business data can be determined according to the business field in the historical business data. The historical to-be-evaluated field refers to the field corresponding to the compliance review of the business data in the reference file. The historical to-be-evaluated field can be at least partially the same as each business field in the historical business data. Optionally, the historical to-be-evaluated field in the reference file can be obtained through text recognition.

[0060] Specifically, the compliance requirement file that matches the historical business data is selected from the compliance requirement file as the reference file, the text in the reference file is recognized, and the historical to-be-evaluated field is determined.

[0061] Optionally, the historical business data is pre-marked with a file identifier of a compliance requirement file that the historical business data needs to meet, and the reference file can be selected from the compliance requirement file according to the file identifier.

[0062] Optionally, the compliance requirement files required to be met by the historical business data of the same business type are usually the same, and the compliance requirement files required to be met by the historical business data of different business types are usually different. Therefore, the reference file can be selected from the compliance requirement file according to the business type to which the historical business data belongs.

[0063] The business type is used to distinguish businesses. The business type to which the historical business data belongs can be determined according to a business field in the historical business data, or according to a pre-set correspondence between a business identifier and a business type. Correspondingly, the compliance requirement file matching the business type is selected as the reference file.

[0064] For example, the compliance requirement file can be subjected to character recognition, and the compliance requirement file with the same character recognition result as the business type is selected as the reference file, where the same character recognition result as the business type can be that the character content in the compliance requirement file is the same as the business type. Alternatively, the business type to which each piece of historical business data belongs is determined according to a pre-set correspondence between a business type and a compliance requirement file identifier required to be met by the business type.

[0065] By determining the reference file from the compliance requirement file according to the business type to which the historical business data belongs, the reference file can be selected, thereby reducing the number of compliance requirement files to be determined for the historical evaluation field, reducing the data operation amount of the historical evaluation field determination process, and further improving the generation efficiency of the audit training sample.

[0066] Optionally, the historical evaluation field in the reference file is determined by extracting a keyword in the reference file and determining the historical evaluation field according to the extraction result.

[0067] Specifically, the reference file is subjected to character recognition to extract a keyword in the reference file, and the keyword matching the extraction result can be found from a pre-set compliance dictionary library as the historical evaluation field. The keyword in the pre-set compliance dictionary library can be set by a technician according to needs or experience.

[0068] It should be noted that the determination of the historical evaluation field by extracting the keyword in the reference file improves the determination efficiency of the historical evaluation field.

[0069] It can be understood that, in order to avoid repeating the extraction of keywords in the same reference file when determining the training sample to be reviewed, the historical evaluation field of different reference files can also be determined in advance in the above manner, and the corresponding historical evaluation field can be obtained when needed.

[0070] In one specific implementation, the extraction of keywords in the reference file and the determination of the historical evaluation field according to the extraction result can include: extracting the keywords of each chapter title in the reference file as the historical evaluation indicator of the corresponding chapter title; and for each chapter title, extracting the keywords of the chapter content corresponding to the chapter title as the historical evaluation field under the corresponding historical evaluation indicator.

[0071] The chapter title refers to the title of the chapter content in the compliance requirement file. At least one chapter title is included in a compliance requirement file. The historical evaluation indicator is used to represent the evaluation field corresponding to the chapter title in the reference file. It can be understood that the historical evaluation field can be classified by the historical evaluation indicator, and the historical evaluation field is used to represent the category to which the historical evaluation field belongs. One chapter title usually corresponds to one historical evaluation indicator.

[0072] The chapter content refers to the content data under the corresponding chapter title in the compliance requirement file. The historical evaluation field under the historical evaluation indicator is used to represent the historical evaluation field contained in the chapter content corresponding to the historical evaluation indicator. The number of historical evaluation fields under the historical evaluation indicator is at least one.

[0073] For example, in a certain compliance requirement file, the historical evaluation indicator of a chapter title is loan ability, and the extracted keywords in the chapter content include loan amount and loan period. Accordingly, the historical evaluation field under the historical evaluation indicator is the loan amount and the loan period.

[0074] Specifically, the text recognition is performed on each chapter title in the reference file, and the recognized keywords are used as the historical evaluation indicator of the corresponding chapter title. For each chapter title, the chapter content corresponding to the chapter title is determined, and the text recognition is performed on the chapter content to determine the keywords of the chapter content. The keywords of the chapter content are used as the historical evaluation field under the historical evaluation indicator.

[0075] By using the keywords of each chapter title in the reference file as the historical evaluation indicator of the corresponding chapter title, and using the keywords of the chapter content corresponding to each chapter title as the historical evaluation field under the corresponding historical evaluation indicator, the hierarchical extraction from the historical evaluation indicator to the historical evaluation field is realized, which facilitates the determination of the historical evaluation field and improves the accuracy and comprehensiveness of the historical evaluation field acquisition.

[0076] S230. Determine the field values ​​corresponding to the historical fields to be evaluated in the historical business data.

[0077] Specifically, for historical fields to be evaluated, the corresponding field values ​​are searched in historical business data. It's understood that historical business data may include some of the field values ​​corresponding to the historical fields to be evaluated. If the historical business data does not include the field values ​​for the historical fields to be evaluated, the corresponding field values ​​can be set to default values. These default values ​​can be set by technical personnel based on needs or experience; for example, the default value can be empty or 0.

[0078] S240. Generate training samples to be reviewed based on the historical fields to be evaluated and their values.

[0079] Specifically, for each piece of historical business data, the field values ​​and default values ​​corresponding to the historical fields to be evaluated can be combined according to the preset order of different historical fields to be evaluated, and the resulting matrix can be used as a training sample to be reviewed.

[0080] S250. Train the pre-built compliance model based on the training samples to be audited and the corresponding compliance result labels.

[0081] The technical solution of this application embodiment reduces the number of reference documents by obtaining the compliance requirement documents corresponding to historical business data as reference documents. By using the historical fields to be evaluated in the reference documents and the field values ​​corresponding to the historical fields to be evaluated in the historical business data, the training samples to be reviewed can be determined, which can reduce the amount of data computation in the process of determining the training samples to be reviewed, thereby improving the efficiency of determining the training samples to be reviewed.

[0082] Based on the above technical solutions, this application also provides an optional embodiment. In this optional embodiment, "determining the training samples to be audited based on compliance requirement documents and historical business data" is specified as follows: "using the compliance requirement documents as reference documents and determining the historical fields to be evaluated in the reference documents; determining the field values ​​corresponding to the historical fields to be evaluated in the historical business data; generating training samples to be audited based on the historical fields to be evaluated and their field values," thereby enriching the mechanism for determining the training samples to be audited. It should be noted that for parts not described in detail in the embodiments of this application, please refer to the relevant descriptions in other embodiments.

[0083] like Figure 3 The compliance model training method shown includes the following specific steps:

[0084] S310, Obtain compliance requirement documents, historical business data, and compliance result tags for historical business data.

[0085] S320, taking the compliance requirement file as a reference file, and determining a history to-be-evaluated field in the reference file.

[0086] Specifically, the compliance requirement file is directly taken as the reference file, and text recognition is performed on the reference file. According to the recognition result, the history to-be-evaluated field in the reference file is determined.

[0087] The determination of the history to-be-evaluated field in the reference file can be understood with reference to the related description of the foregoing embodiments, which will not be repeated here.

[0088] S330, determining a field value corresponding to the history to-be-evaluated field in the history business data.

[0089] Specifically, the field value corresponding to the history to-be-evaluated field is searched for in the history business data. In this embodiment, the history to-be-evaluated field without corresponding history business data can be ignored, or the field value of the corresponding history to-be-evaluated field can be set as a default value, for example, the default value can be a null value or a 0 value.

[0090] Optionally, the determination of the field value corresponding to the history to-be-evaluated field in the history business data includes: taking a business field in the history business data that matches the history to-be-evaluated field as a reference business field, and determining a field value of the reference business field in the history business data; and correspondingly, generating a to-be-audited training sample according to the history to-be-evaluated field and the field value of the history to-be-evaluated field includes: generating the to-be-audited training sample according to the reference business field and the field value of the reference business field.

[0091] The reference business field refers to a business field in the history business data that is the same as the history to-be-evaluated field. Specifically, the business fields in the history business data are compared with the history to-be-evaluated field, and the business field in the history business data that is the same as the history to-be-evaluated field is taken as the reference business field. The value corresponding to the reference business field in the history business data is searched for as the field value of the reference business field. The field value of the history to-be-evaluated field that is not searched for in the history business data is set as a default value. The reference business field, the field value of the reference business field, and the history to-be-evaluated field that is not searched for and the corresponding default value can be combined according to a preset arrangement order of the history to-be-evaluated fields, and the matrix obtained by the combination can be taken as the to-be-audited training sample.

[0092] By determining the reference business field in the business field in the history business data, the field value of the reference business field is determined, which can generate the corresponding to-be-audited training sample by using the same processing mode for different history business data, and realizes the standardization and universality of the to-be-audited training sample generation process.

[0093] S340. Generate training samples to be reviewed based on the historical fields to be evaluated and their values.

[0094] Specifically, the historical fields to be evaluated and their values ​​can be combined to form a matrix, which can then be used as training samples for review.

[0095] S350. Train the pre-built compliance model based on the training samples to be audited and the corresponding compliance result labels.

[0096] The technical solution of this application embodiment uses compliance requirement documents as reference documents to determine historical fields to be evaluated in the reference documents, avoiding omissions in the reference documents and thus increasing the scope of determination of historical fields to be evaluated, improving the richness and comprehensiveness of these fields. Simultaneously, the same method is used to generate training samples to be audited for each compliance requirement document and historical business data, ensuring the standardization and universality of the training sample generation process.

[0097] The training method for the compliance model has been explained in detail above. The following section will describe in detail the process of using the compliance model.

[0098] Figure 4 This is a flowchart illustrating a compliance assessment method provided in an embodiment of this application. This embodiment is applicable to situations where various compliance models are used. The method can be executed by a compliance assessment device, which can be implemented in software and / or hardware and can be configured in an electronic device. It should be noted that the electronic device executing the compliance assessment method may be the same as or different from the electronic device used for the aforementioned compliance model training method; this application does not impose any limitations on this.

[0099] like Figure 4 The compliance assessment method shown includes the following steps:

[0100] S410. Determine the pending business data and the pending compliance requirements documents for the pending business data.

[0101] A pending business transaction refers to a business transaction that requires review. In a specific example, a pending business transaction could be a loan transaction. Pending business transaction data refers to the data provided by the business initiator when initiating a pending business transaction. Pending business transaction data may include business fields and their corresponding values. Pending compliance requirement documents for pending business data refer to the compliance requirement documents required to review the pending business data. There must be at least one pending compliance requirement document for pending business data. The compliance requirement documents that the pending business data needs to meet can be used as the pending review documents for the pending business data.

[0102] Optionally, the to-be-audited business data of the to-be-audited business is acquired, and the compliance requirement file checked by the to-be-audited business when initiated is taken as the to-be-audited compliance requirement file.

[0103] For example, the to-be-audited business data of the to-be-audited business is determined according to the initiated to-be-audited business. When the to-be-audited business is initiated, a list of compliance requirement files can be provided through an interactive interface, and the business initiator checks the compliance requirement files on the interactive interface. The business initiator can select the to-be-audited compliance requirement file from the compliance requirement files by checking, thereby realizing on-demand selection of the to-be-audited compliance requirement file and improving the flexibility and convenience of selection of the to-be-audited compliance requirement file.

[0104] Alternatively, the to-be-audited business data of the to-be-audited business is acquired, and the to-be-audited compliance requirement file is selected from the compliance requirement files according to the business type to which the to-be-audited business data belongs.

[0105] For example, the business type is determined according to the to-be-audited business data, the compliance requirement files are screened according to the business type, the compliance requirement file matched with the business type is selected as the to-be-audited compliance requirement file. Specifically, the business type to which the to-be-audited business belongs is taken as the to-be-audited business type, and the compliance requirement file corresponding to the to-be-audited business type is selected as the to-be-audited compliance requirement file.

[0106] By automatically determining the to-be-audited compliance requirement file according to the business type to which the to-be-audited business belongs, the business initiator does not need to manually select, thereby improving the accuracy and convenience of the to-be-audited compliance requirement file.

[0107] S420, determining the to-be-audited prediction sample according to the to-be-audited compliance requirement file and the to-be-audited business data.

[0108] The to-be-audited prediction sample is the input data of the trained compliance model, and the number of the to-be-audited prediction sample is at least one. Specifically, according to the to-be-audited compliance requirement file and the to-be-audited business data, the associated data of the to-be-audited compliance requirement file in the to-be-audited business data can be determined, and thus the to-be-audited prediction sample can be determined according to the associated data.

[0109] S430, inputting the to-be-audited prediction sample into the trained compliance model to obtain the compliance result.

[0110] The compliance model is trained based on any of the above compliance model training methods.

[0111] The compliance result is used to represent an audit result obtained after the to-be-audited prediction sample is audited, that is, an audit result obtained after the to-be-audited business data is audited. The compliance result can be compliance or non-compliance.

[0112] Specifically, the to-be-audited prediction sample is taken as input data of the trained compliance model, the trained compliance model is used, and the compliance result is determined according to the model output result.

[0113] The model output result can be a binary classification result of compliance or non-compliance; or the model output result can be a compliance probability of the to-be-audited prediction sample; if the compliance probability is greater than a first preset threshold, the to-be-audited prediction sample is determined to be compliant; if the compliance probability is less than a second preset threshold, the to-be-audited prediction sample is determined to be non-compliant. The first preset probability is not less than the second preset probability; the first preset probability and the second preset probability can be set by technical personnel according to needs and experience values, or repeatedly determined through a large number of experiments.

[0114] In an implementable manner, the to-be-audited compliance requirement file can be updated according to actual conditions, so that after the compliance result is obtained, the compliance result can be verified by manual auditing, and the verification result is taken as the final compliance result.

[0115] The technical scheme of the embodiment of the application determines the to-be-audited prediction sample through the to-be-audited business data of the to-be-audited business and the to-be-audited compliance requirement file of the to-be-audited business data, inputs the to-be-audited prediction sample into the trained compliance model, and obtains the compliance result, without summarizing and refining the rules audited by the compliance requirement file, thereby improving the convenience of the compliance review process. At the same time, the above compliance review method can simultaneously implement compliance review of multiple compliance requirement files, has a wider application range, and is more universal. At the same time, the trained compliance model based on artificial intelligence has better compliance distinguishing ability, reduces the influence of subjectivity, and improves the accuracy of the compliance review result.

[0116] In an optional embodiment, if the compliance model introduces the initiator category of the business initiator as a to-be-audited training sample in the training process, the model is trained. Correspondingly, in the model use stage, the initiator category of the business initiator needs to be introduced as a to-be-audited prediction sample.

[0117] For example, the to-be-audited prediction sample is determined according to the to-be-audited compliance requirement file and the to-be-audited business data, which can include: obtaining the initiator category of the business initiator corresponding to the to-be-audited business; determining the to-be-audited prediction sample according to the initiator category corresponding to the to-be-audited business, the to-be-audited compliance requirement file, and the to-be-audited business data.

[0118] Specifically, the compliance requirement file corresponding to the initiator category of the business initiator corresponding to the to-be-audited business can be selected from a large number of compliance requirement files, and the to-be-audited prediction sample can be determined according to the associated data of the to-be-audited business data corresponding to the compliance requirement file corresponding to the initiator category.

[0119] By introducing the initiator category of the business initiator corresponding to the to-be-audited business data into the to-be-audited prediction sample when the model is used for compliance review, the compliance review can be performed in a targeted manner for business initiators of different categories, and the flexibility and accuracy of the determination of the to-be-audited prediction sample are improved.

[0120] On the basis of the above technical solutions, the present application further provides an optional embodiment, in which the "determining the to-be-audited prediction sample according to the to-be-audited compliance requirement file and the to-be-audited business data" is specified as "determining a to-be-evaluated field in the to-be-audited compliance requirement file; determining a field value corresponding to the to-be-evaluated field in the to-be-audited business data; and generating the to-be-audited prediction sample according to the to-be-evaluated field and the field value of the to-be-evaluated field", so as to improve the generation mechanism of the to-be-audited prediction sample in the model use stage. It should be noted that the parts not described in detail in the embodiments of the present application can be referred to the related descriptions of other embodiments.

[0121] As shown in Figure 5 The compliance judgment method comprises the following specific steps:

[0122] S510, determining to-be-audited business data of a to-be-audited business and a to-be-audited compliance requirement file of the to-be-audited business data.

[0123] S520, determining a to-be-evaluated field in the to-be-audited compliance requirement file.

[0124] The to-be-evaluated field refers to a field in the to-be-audited compliance requirement file according to which the to-be-audited business data is subjected to compliance review. The to-be-evaluated field can be at least partially the same as each business field in the to-be-audited business data. For example, for the to-be-audited compliance requirement file, the to-be-evaluated field in the to-be-audited compliance requirement file can be obtained by means of character recognition.

[0125] In an optional embodiment, the keywords in the to-be-audited compliance requirement file can be extracted, and the to-be-evaluated field can be determined according to the extraction result.

[0126] Specifically, the keywords in the to-be-audited compliance requirement file can be extracted, and the keywords matching the extraction result can be searched from a preset compliance dictionary library as the to-be-evaluated field. The keywords in the preset compliance dictionary library can be set by technical personnel according to needs or experience.

[0127] It should be noted that by extracting the keywords in the to-be-audited file, the automation of the to-be-evaluated field is determined, and the determination efficiency of the to-be-evaluated field is improved.

[0128] It can be understood that, in order to avoid repeatedly extracting the keywords in the same to-be-audited compliance requirement file when determining the to-be-predicted sample, the to-be-evaluated fields of different compliance requirement files can also be determined in advance in the above manner, and the corresponding to-be-evaluated fields can be obtained when needed.

[0129] In an optional embodiment, extracting the keywords in the to-be-audited compliance requirement file and determining the to-be-evaluated field according to the extraction result can include: extracting the keywords of each chapter title in the to-be-audited compliance requirement file as the to-be-evaluated indicators of the corresponding chapter title; for each chapter title, extracting the keywords of the chapter content corresponding to the chapter title as the to-be-evaluated field under the corresponding to-be-evaluated indicator.

[0130] The to-be-evaluated indicator is used to represent the evaluation field corresponding to the chapter title in the to-be-audited compliance requirement file, and it can also be understood that the to-be-evaluated field can be classified by the to-be-evaluated indicator, and the to-be-evaluated indicator can be used to represent the category to which the to-be-evaluated field belongs. A chapter title usually corresponds to a to-be-evaluated indicator. The to-be-evaluated field under the to-be-evaluated indicator is used to represent the to-be-evaluated field contained in the chapter where the to-be-evaluated indicator is located. The number of to-be-evaluated fields under the to-be-evaluated indicator is at least one.

[0131] Specifically, text recognition is performed on each chapter title in the to-be-audited compliance requirement file to obtain the keywords of each chapter title, and the keywords of each chapter title are used as the to-be-evaluated indicators of the corresponding chapter. For each chapter title, determine the chapter content corresponding to the chapter title, and perform text recognition on the chapter content to determine the keywords of the chapter content, and the keywords of the chapter content are used as historical to-be-evaluated fields under historical to-be-evaluated indicators.

[0132] By using the keywords of each chapter title in the to-be-audited compliance requirement file as the to-be-evaluated indicators of the corresponding chapter title, and using the keywords of the chapter content corresponding to each chapter title as the historical to-be-evaluated fields under the corresponding to-be-evaluated indicators, the to-be-evaluated indicators are extracted layer by layer, which facilitates the determination of the to-be-evaluated field and improves the accuracy and comprehensiveness of the to-be-evaluated field.

[0133] S530, determine the field value corresponding to the to-be-evaluated field in the to-be-audited business data.

[0134] Specifically, for the field to be evaluated, the corresponding field value is searched in the business data to be reviewed. It's understood that the business data to be reviewed may only include a portion of the field values ​​corresponding to the field to be evaluated. If the field value to be evaluated is not included in the business data to be reviewed, the corresponding field value can be set to a default value. This default value can be set by technical personnel based on needs or experience; for example, the default value can be empty or 0.

[0135] S540. Generate a prediction sample to be reviewed based on the field to be evaluated and its value.

[0136] Specifically, the field values ​​and default values ​​corresponding to the fields to be evaluated can be combined according to the preset order of different historical fields to be evaluated, and the resulting matrix can be used as the prediction sample to be reviewed.

[0137] S550. Input the predicted sample to be audited into the trained compliance model to obtain the compliance result; wherein, the compliance model is trained based on any of the compliance model training methods mentioned above.

[0138] The technical solution of this application embodiment automatically determines the field values ​​of the fields to be evaluated in the compliance requirements document to be reviewed, thereby improving the efficiency and accuracy of the determination of the field values ​​corresponding to the fields to be evaluated, thus improving the accuracy of the determination of the predicted samples to be reviewed, and further improving the accuracy of the compliance review results of the business data to be reviewed.

[0139] As an implementation of the above-mentioned compliance model training methods, this application also provides an optional embodiment of an execution device for implementing the above-mentioned compliance model training methods.

[0140] The compliance model training apparatus provided in this optional embodiment is suitable for training various compliance models. The apparatus can be implemented in software and / or hardware and can be configured in an electronic device.

[0141] like Figure 6 The compliance model training device shown specifically includes: a data acquisition module 601, a training sample determination module 602, and a training module 603. Among them,

[0142] Data acquisition module 601 is used to acquire compliance requirement documents, historical business data, and compliance result tags of historical business data;

[0143] The training sample determination module 602 is used to determine the training samples to be audited based on compliance requirement documents and historical business data.

[0144] The training module 603 is configured to train the pre-constructed compliance model according to the training sample to be audited and the corresponding compliance result label.

[0145] The technical scheme of the embodiment of the application determines the training sample to be audited according to the compliance requirement file and the historical business data, trains the pre-constructed compliance model according to the compliance result label of the training sample to be audited and the historical business data, so that the trained compliance model has the compliance review capability of multiple compliance requirement files, thereby achieving higher accuracy in subsequent compliance review using the compliance model, and achieving compliance review of multiple compliance requirement files, and having a wider application range. In addition, the compliance review process is more convenient without summarizing and refining the audited rules.

[0146] Optionally, the training sample determination module 602 comprises:

[0147] The first historical field determination unit is configured to take the historical business data corresponding to the compliance requirement file as a reference file, and determine a historical evaluation field in the reference file;

[0148] The first historical field value determination unit is configured to determine a field value corresponding to the historical evaluation field in the historical business data;

[0149] The first training sample determination unit is configured to generate the training sample to be audited according to the historical evaluation field and the field value of the historical evaluation field.

[0150] Optionally, the first historical field determination unit is specifically configured to:

[0151] According to the business type to which the historical business data belongs, the reference file is selected from the compliance requirement file.

[0152] Optionally, the training sample determination module 602 comprises:

[0153] The second historical field determination unit is configured to take the compliance requirement file as a reference file, and determine a historical evaluation field in the reference file;

[0154] The second historical field value determination unit is configured to determine a field value corresponding to the historical evaluation field in the historical business data;

[0155] The second training sample determination unit is configured to generate the training sample to be audited according to the historical evaluation field and the field value of the historical evaluation field.

[0156] Optionally, the second historical field value determination unit is specifically configured to:

[0157] The historical business data is matched with the historical evaluation field, and a reference business field in the historical business data is determined as a reference business field, and a field value of the reference business field in the historical business data is determined.

[0158] Correspondingly, the second training sample determination unit is specifically configured to:

[0159] generate the training sample to be audited according to the reference business field and the field value of the reference business field.

[0160] Optionally, the first historical field determination unit or the second historical field determination unit is specifically configured to:

[0161] extract the keywords in the reference file, and determine the historical evaluation field according to the extraction result.

[0162] Optionally, the first historical field determination unit or the second historical field determination unit comprises:

[0163] the historical index determination subunit is configured to extract the keywords of the chapter titles in the reference file as the historical evaluation indexes of the corresponding chapter titles;

[0164] the historical field determination subunit is configured to extract the keywords of the chapter contents corresponding to the chapter titles as the historical evaluation fields under the corresponding historical evaluation indexes.

[0165] Optionally, the training sample determination module 602 comprises:

[0166] the historical initiator category determination unit is configured to obtain the initiator category of the business initiator corresponding to the historical business data;

[0167] the third training sample determination unit is configured to determine the training sample to be audited according to the initiator category corresponding to the historical business data, the compliance requirement file and the historical business data.

[0168] The above compliance model training device can execute the compliance model training method provided in any embodiment of the present application, and has the corresponding function modules and beneficial effects of executing each compliance model training method.

[0169] As an implementation of each of the above compliance judgment methods, the present application further provides an optional embodiment of an execution device for implementing each of the above compliance judgment methods. The compliance judgment device provided by the optional embodiment is suitable for the case of using a compliance model for compliance review. The device can be implemented in the form of software and / or hardware, and can be configured in an electronic device.

[0170] As shown in the compliance judgment device, Figure 7 specifically comprises a compliance requirement file determination module 701, a prediction sample determination module 702 and a compliance result determination module 703. Among them,

[0171] The compliance requirement file determination module 701 is configured to determine the to-be-audited business data of the to-be-audited business and the to-be-audited compliance requirement file of the to-be-audited business data;

[0172] The prediction sample determination module 702 is configured to determine the to-be-audited prediction sample according to the to-be-audited compliance requirement file and the to-be-audited business data.

[0173] The compliance result determination module 703 is configured to input the to-be-audited prediction sample into the trained compliance model to obtain the compliance result, wherein the compliance model is trained based on the compliance model training apparatus.

[0174] The technical scheme of the embodiment of the present application determines the to-be-audited prediction sample through the to-be-audited business data of the to-be-audited business and the to-be-audited compliance requirement file of the to-be-audited business data, inputs the to-be-audited prediction sample into the trained compliance model to obtain the compliance result, and does not need to summarize and refine the rules audited by the compliance requirement file, thereby improving the convenience of the compliance review process. Meanwhile, the compliance review method can simultaneously implement compliance review of multiple compliance requirement files, has a wider application range, and is more universal. Meanwhile, the compliance model is trained based on the artificial intelligence method, the trained compliance model has better compliance recognition ability, the influence of subjectivity is reduced, and the accuracy of the compliance review result is improved.

[0175] Optionally, the prediction sample determination module 702 comprises:

[0176] The field determination unit is configured to determine the to-be-evaluated field in the to-be-audited compliance requirement file.

[0177] The field value determination unit is configured to determine the field value corresponding to the to-be-evaluated field in the to-be-audited business data.

[0178] The prediction sample determination unit is configured to generate the to-be-audited prediction sample according to the to-be-evaluated field and the field value of the to-be-evaluated field.

[0179] Optionally, the field determination unit is specifically configured to:

[0180] extract the keywords in the to-be-audited compliance requirement file, and determine the to-be-evaluated field according to the extraction result.

[0181] Optionally, the field determination unit comprises:

[0182] The index determination subunit is configured to extract the keywords of the chapter titles in the to-be-audited compliance requirement file as the to-be-evaluated index of the corresponding chapter titles.

[0183] The field determination subunit is configured to extract the keywords of the chapter content corresponding to the chapter title as the to-be-evaluated field under the corresponding to-be-evaluated index for each chapter title.

[0184] Optional, the compliance requirements documentation determination module 701 includes:

[0185] The business data acquisition unit is used to acquire the business data to be reviewed for the business to be reviewed.

[0186] The first compliance requirement document determination unit is used to determine the compliance requirement documents selected when the business to be audited was initiated as the compliance requirement documents to be audited; and / or,

[0187] The second compliance requirement document determination unit is used to select the compliance requirement document to be audited from the compliance requirement documents based on the business type to which the business data to be audited belongs.

[0188] Optionally, the prediction sample determination module 702 includes:

[0189] The initiator category determination unit is used to obtain the initiator category of the business initiator corresponding to the business to be reviewed;

[0190] The prediction sample determination unit is used to determine the prediction sample to be audited based on the initiator category, compliance requirement documents, and business data of the business to be audited.

[0191] The aforementioned compliance assessment device can execute the compliance assessment method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the compliance assessment method.

[0192] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 8 A block diagram is shown of an exemplary electronic device 800 suitable for implementing embodiments of the present application. Figure 8 The electronic device 800 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0193] like Figure 8 As shown, the electronic device 800 is presented in the form of a general-purpose computing device. The components of the electronic device 800 may include, but are not limited to: one or more processors or processing units 801, system memory 802, and bus 803 connecting different system components (including system memory 802 and processing unit 801).

[0194] Bus 803 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0195] Electronic device 800 typically includes a variety of computer system readable media. Such media can be any available media that is located either internally or externally to electronic device 800, such as volatile and non-volatile media, removable and non-removable media.

[0196] System memory 802 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) 804 and / or cache memory 805. Electronic device 800 can further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 806 can be provided for reading from and writing to a non-removable, non-volatile magnetic media (not shown and typically called a "hard drive"). Figure 8 not shown, typically referred to as a "hard disk drive", for reading from and writing to non-removable, non-volatile magnetic media (not shown). Although Figure 8 In alternative embodiments, a magnetic hard disk drive (not shown), a solid state drive (not shown), or a optical disk drive (not shown) can be used for inputting or outputting data and / or computer instructions to or from electronic device 800. These and other variations of electronic device 800 are contemplated by the present disclosure. Although custom has been to name certain components of the electronic device 800 after articles of clothing, the naming of the component is exemplary only and should not be construed as a limitation of the electronic device 800. In fact, any device that includes a processing system can be used with the present disclosure.

[0197] Program / utility 808 having a set (at least one) of program modules 807, can be stored in memory 802 by way of example, and not limitation, as well as an operating system, one or more application programs, other program modules, and program data, each or some combination thereof, can include implementation of the network environment. Program modules 807 generally carry out the functions and / or methodologies of embodiments of the present disclosure.

[0198] Electronic device 800 can also communicate with one or more external devices 809 such as a keyboard or a pointing device, displays 810, etc.; one or more devices that enable a user to interact with electronic device 800; and / or one or more devices that enable electronic device 800 to communicate with one or more other computing devices. Such communication can be via input / output (I / O) interfaces 811. Similarly, such communication can be via network adapter 812. Network adapter 812 can communicate with the other Figure 8Other hardware and / or software modules can be used in conjunction with electronic device 800, as indicated by the dashed line in FIG. 8. Such hardware and / or software modules include, but are not limited to, microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0199] The processing unit 801 performs various function applications and data processing by running programs stored in the system memory 802, such as implementing the compliance model training method or the compliance judgment method provided in the embodiments of the present application.

[0200] The embodiments of the present application also provide a storage medium containing computer executable instructions, and a computer program is stored on the storage medium. The program is executed by a processor to implement the compliance model training method provided in the embodiments of the present application, which includes: obtaining a compliance requirement file, historical business data and a compliance result label of the historical business data; determining a to-be-audited training sample according to the compliance requirement file and the historical business data; and training a pre-constructed compliance model according to the to-be-audited training sample and the corresponding compliance result label.

[0201] The embodiments of the present application also provide a storage medium containing computer executable instructions, and a computer program is stored on the storage medium. The program is executed by a processor to implement the compliance judgment method provided in the embodiments of the present application, which includes: determining to-be-audited business data of a to-be-audited business and a to-be-audited compliance requirement file of the to-be-audited business data; determining a to-be-audited prediction sample according to the to-be-audited compliance requirement file and the to-be-audited business data; and inputting the to-be-audited prediction sample into a trained compliance model to obtain a compliance result; wherein the compliance model is trained based on the compliance model training method provided in any of the embodiments of the present application.

[0202] The computer storage medium of the embodiments of the present application can adopt any combination of one or more computer readable media. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination thereof. More specific examples (non-exhaustive list) of the computer readable storage medium include: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus.

[0203] A computer readable signal medium can include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated signal can take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium can be any computer readable medium that can be involved in

[0204] The code can be transmitted in any form, including but not limited to radio frequency, light, electrical, or the like, or any suitable combination thereof.

[0205] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0206] The embodiments of the present application further provide a computer program product, comprising a computer program which, when executed by a processor, implements the compliance model training method or the compliance judgment method provided by any of the embodiments of the present application.

[0207] Computer program product, in the implementation of the process, can be written in one or more programming languages or combinations thereof for computer program code for performing the operations of the present application, programming language includes object-oriented programming language, such as Java, Smalltalk, C++, also includes conventional procedural programming language, such as "C" language or similar programming language. Program code can be completely executed on a user computer, partially executed on a user computer, as an independent software package, partially on a user computer and partially on a remote computer, or completely on a remote computer or server. In the case of remote computer, the remote computer can be connected to the user computer through any kind of network, including local area network (LAN) or wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0208] Note that the above is only the preferred embodiment of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments herein, and those skilled in the art can make various obvious changes, readjustments and substitutions without departing from the scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A compliance model training method, characterized in that, include: Obtain compliance requirement documents, historical business data, and compliance result tags for the historical business data; Based on the aforementioned compliance requirements documents and the aforementioned historical business data, the training samples to be audited were determined; The pre-built compliance model is trained based on the training samples to be audited and the corresponding compliance result labels. The step of determining the training samples to be audited based on the compliance requirement documents and the historical business data includes: The compliance requirement document corresponding to the historical business data is used as a reference document, and the historical fields to be evaluated in the reference document are determined; the field values ​​corresponding to the historical fields to be evaluated in the historical business data are determined; the business fields in the historical business data that match the historical fields to be evaluated are used as reference business fields, and the field values ​​of the reference business fields in the historical business data are determined; the training samples to be reviewed are generated based on the reference business fields and the field values ​​of the reference business fields. The step of determining the training samples to be audited based on the compliance requirement documents and the historical business data further includes: Obtain the initiator category of the business initiator corresponding to the historical business data; Based on the initiator category corresponding to the historical business data, the compliance requirement document, and the historical business data, the training samples to be audited are determined; the initiator category is the category to which the business initiator belongs; the business initiator is an enterprise, organization, or individual; The determination of the historical fields to be evaluated in the reference file includes: Extract keywords from the reference file and determine the historical fields to be evaluated based on the extraction results; The step of extracting keywords from the reference file and determining the historical fields to be evaluated based on the extraction results includes: Extract keywords from the chapter titles of the reference files and use them as historical evaluation indicators for the corresponding chapter titles; For each chapter title, extract the keywords corresponding to the chapter content and use them as historical evaluation fields under the corresponding historical evaluation indicators.

2. The method according to claim 1, characterized in that, The phrase "using the compliance requirement document corresponding to the historical business data as a reference document" includes: Based on the business type to which the historical business data belongs, select the reference document from the compliance requirements document.

3. A compliance assessment method, characterized in that, include: Identify the business data to be audited and the compliance requirement documents for the business data to be audited; Based on the compliance requirement documents to be audited and the business data to be audited, a sample of samples to be audited is determined; The predicted sample to be audited is input into the trained compliance model to obtain the compliance result; wherein the compliance model is trained based on the method described in any one of claims 1-2; The step of determining the audit prediction sample based on the compliance requirement document to be audited and the business data to be audited includes: Identify the fields to be evaluated in the compliance requirements document to be reviewed; Determine the field value corresponding to the field to be evaluated in the business data to be reviewed; Based on the field to be evaluated and its value, generate the prediction sample to be reviewed; The step of determining the audit prediction sample based on the compliance requirement document to be audited and the business data to be audited further includes: Obtain the initiator category of the business initiator corresponding to the business to be reviewed; Based on the initiator category corresponding to the business to be audited, the compliance requirement document to be audited, and the business data to be audited, a predicted sample to be audited is determined; the initiator category is the category to which the business initiator belongs; the business initiator is an enterprise, organization, or individual; The process of determining the fields to be evaluated in the compliance requirements document to be audited includes: Extract keywords from the compliance requirements document to be reviewed, and determine the fields to be evaluated based on the extraction results; The step of extracting keywords from the compliance requirements document to be reviewed and determining the fields to be evaluated based on the extraction results includes: Extract keywords from the chapter titles of the compliance requirements document to be reviewed, and use them as evaluation indicators for the corresponding chapter titles; For each chapter title, extract the keywords corresponding to the chapter content and use them as the evaluation fields under the corresponding evaluation indicators.

4. The method according to claim 3, characterized in that, The determination of the pending business data and the pending compliance requirement document for the pending business data includes: Retrieve the pending business data for the pending business; The compliance requirement document selected when initiating the business to be reviewed shall be used as the compliance requirement document to be reviewed; and / or, Based on the business type to which the business data to be audited belongs, select the compliance requirement document to be audited from the compliance requirement documents.

5. A compliant model training device, characterized in that, include: The data acquisition module is used to acquire compliance requirement documents, historical business data, and compliance result tags of the historical business data; The training sample determination module is used to determine the training samples to be audited based on the compliance requirement document and the historical business data. The training module is used to train the pre-built compliance model based on the training samples to be audited and the corresponding compliance result labels. The training sample determination module includes: The second historical field determination unit is used to use the compliance requirement document as a reference document and determine the historical fields to be evaluated in the reference document. The second historical field value determination unit is used to determine the field value corresponding to the historical field to be evaluated in the historical business data; The second training sample determination unit is used to generate the training sample to be reviewed based on the historical field to be evaluated and the field value of the historical field to be evaluated. The training sample determination module includes: The historical initiator category determination unit is used to obtain the initiator category of the business initiator corresponding to the historical business data; The third training sample determination unit is used to determine the training samples to be reviewed based on the initiator category corresponding to the historical business data, the compliance requirement document, and the historical business data; the initiator category is the category to which the business initiator belongs; the business initiator is an enterprise, organization, or individual; The training sample determination module includes: The first historical field determination unit is used to take the compliance requirement document corresponding to the historical business data as a reference document and determine the historical fields to be evaluated in the reference document. The first historical field value determination unit is used to determine the field value corresponding to the historical field to be evaluated in the historical business data; The first training sample determination unit is used to generate the training sample to be reviewed based on the historical field to be evaluated and the field value of the historical field to be evaluated. The first historical field determination unit or the second historical field determination unit is specifically used for: Extract keywords from the reference file and determine the historical fields to be evaluated based on the extraction results; The first historical field determination unit or the second historical field determination unit includes: The historical indicator determination sub-unit is used to extract keywords from the chapter titles of the reference documents, which are then used as historical indicators to be evaluated for the corresponding chapter titles. The historical field determines the sub-unit, which is used to extract the keywords of the chapter content corresponding to each chapter title, and use them as the historical evaluation field under the corresponding historical evaluation indicator.

6. A compliance judgment device, characterized in that, include: The compliance requirement document determination module is used to determine the business data to be audited and the compliance requirement document to be audited for the business data to be audited. The prediction sample determination module determines the prediction sample to be audited based on the compliance requirement document to be audited and the business data to be audited. The compliance result determination module is used to input the predicted sample to be audited into a trained compliance model to obtain a compliance result; wherein, the compliance model is trained based on the device described in claim 5; The prediction sample determination module includes: A field determination unit is used to determine the fields to be evaluated in the compliance requirements document to be reviewed. A field value determination unit is used to determine the field value corresponding to the field to be evaluated in the business data to be reviewed; The prediction sample determination unit is used to generate the prediction sample to be reviewed based on the field to be evaluated and the field value of the field to be evaluated. The prediction sample determination module further includes: The initiator category determination unit is used to obtain the initiator category of the business initiator corresponding to the business to be reviewed; The prediction sample determination unit is used to determine the prediction sample to be reviewed based on the initiator category corresponding to the business to be reviewed, the compliance requirement document to be reviewed, and the business data to be reviewed; the initiator category is the category to which the business initiator belongs; the business initiator is an enterprise, organization, or individual; The field determination unit is specifically used for: Extract keywords from the compliance requirements documents to be reviewed, and determine the fields to be evaluated based on the extraction results; The field determination unit includes: The indicator determination sub-unit is used to extract keywords from the chapter titles of the compliance requirements document to be reviewed, and use them as the evaluation indicators for the corresponding chapter titles. The field determines the sub-unit, which is used to extract keywords from the content of the chapter corresponding to each chapter title, and use them as the fields to be evaluated under the corresponding evaluation indicators.

7. The apparatus according to claim 6, characterized in that, The compliance requirement document determination module includes: The business data acquisition unit is used to acquire the business data to be reviewed for the business to be reviewed. The first compliance requirement document determination unit is used to determine the compliance requirement document selected when the business to be audited was initiated as the compliance requirement document to be audited; and / or, The second compliance requirement document determination unit is used to select the compliance requirement document to be audited from the compliance requirement documents according to the business type to which the business data to be audited belongs.

8. An electronic device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the compliance model training method as described in any one of claims 1-2, or the compliance judgment method as described in any one of claims 3-4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the compliance model training method as described in any one of claims 1-2, or the compliance judgment method as described in any one of claims 3-4.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the compliance model training method as described in any one of claims 1-2, or the compliance judgment method as described in any one of claims 3-4.

Images