A financial business audit processing method based on large model

Abstract

The present invention relates to the field of financial business audit processing, and in particular to a financial business audit processing method based on a large model, including obtaining business data to be audited sent by several sending ends, and analyzing the business data to be audited of each sending end to extract business feature data of the business data to be audited; obtaining node feature data of several nodes in a target database; randomly combining the business feature data and the node feature data to generate a disguise mask; sending the business data to be audited and the disguise mask to a supervisory end; when the supervisory end completes sending the business data to be audited, the supervisory end allocates the business data to be audited received by the target database to the target node; using the business data to be audited and the audit results in the historical supervision process of the supervisory end on several business data to be audited as an audit data set to train a business large model, and auditing subsequent businesses to be audited based on the trained large model. The present invention improves the efficiency of financial business audit processing.

Description

Technical Field

[0001] The present invention relates to the field of financial business audit processing, and in particular to a financial business audit processing method based on a large model. Background Art

[0002] With the rapid development of artificial intelligence technology, especially the emergence of big models, the financial industry has ushered in new opportunities for digital transformation. Big model technology, with its powerful natural language processing and knowledge reasoning capabilities, provides a new solution for financial business audit processing. These technologies can not only improve the efficiency and accuracy of audits, but also expand the scenarios of financial services, optimize customer experience, and thus enhance the core competitiveness of financial institutions.

[0003] Chinese patent application publication number: CN115936628A discloses a data audit method, device, equipment, storage medium and computer program product, including responding to an application request of a to-be-audited object for a target business, obtaining a target audit template corresponding to the target business, obtaining the to-be-audited data type corresponding to the target business, the data source information of the to-be-audited data type and the audit model of the to-be-audited data type from the target audit template, obtaining the to-be-audited data of the to-be-audited object from the data source corresponding to the data source information, the to-be-audited data corresponding to the to-be-audited data type, and performing audit processing on the to-be-audited data according to the audit model to obtain an audit result corresponding to the application request. However, the prior art has the following problems:

[0004] The existing technology has the problem of misjudgment or missed judgment, resulting in low accuracy of financial business review and processing, thereby causing low efficiency of financial business review and processing. Summary of the invention

[0005] To this end, the present invention provides a financial business review and processing method based on a large model to overcome the problem of low efficiency of financial business review and processing in the prior art.

[0006] To achieve the above object, the present invention provides a financial business review processing method based on a large model, comprising:

[0007] Acquire the business data to be reviewed sent by several sending ends, and analyze the business data to be reviewed of each sending end to extract business characteristic data of the business data to be reviewed;

[0008] Obtain node feature data of several nodes in the target database;

[0009] Randomly combining the service characteristic data and the node characteristic data to generate a camouflage mask;

[0010] Sending the pending business data and the disguised mask to a supervisory end;

[0011] When the sending end completes sending the business data to be reviewed, the supervisory end distributes the business data to be reviewed received by the target database to the target node and determines whether the distribution process of the business data to be reviewed has been tampered with;

[0012] The business data to be reviewed and the review results in the historical supervision process of the supervision end on a number of business data to be reviewed are used as a review data set to train the business big model, and the subsequent business data to be reviewed are reviewed based on the trained big model.

[0013] Furthermore, under the condition of obtaining the business data to be reviewed sent by several sending ends, based on the comparison result that the importance evaluation value of the business data to be reviewed is less than or equal to the preset importance evaluation value, it is determined to extract the characteristic data of the business data to be reviewed based on the data structure change value of the business data to be reviewed, so as to generate business characteristic data.

[0014] Furthermore, under the condition of obtaining business data to be reviewed sent by several sending ends, based on the comparison result that the importance evaluation value of the business data to be reviewed is greater than the preset importance evaluation value, it is determined to extract the characteristic data of the business data to be reviewed based on the data property change value of the business data to be reviewed, so as to generate business characteristic data.

[0015] Furthermore, under the condition that the generation of business characteristic data is completed, based on the comparison result that the comprehensive evaluation value of the business data to be reviewed is less than or equal to the preset comprehensive evaluation value, it is determined that the business data to be reviewed received from the regulatory end is extracted from the data segment with the same characteristics as the business data to be reviewed received from the target database, and the characteristics include changes in data structure and changes in data properties.

[0016] Furthermore, under the condition that the generation of business characteristic data is completed, based on the comparison result that the comprehensive evaluation value of the business data to be reviewed is greater than the preset comprehensive evaluation value, it is determined to extract all data from the business data to be reviewed received from the supervision end and the business data to be reviewed received from the target database.

[0017] Furthermore, under the condition that the data segments with the same characteristics are extracted from the business data to be reviewed received from the regulatory end and the business data to be reviewed received by the target database, it is determined that the process of allocating the business data to be reviewed to the target node has been tampered with based on the inconsistent comparison result between the data volume of the data segments extracted from the business data to be reviewed received by the regulatory end and the data volume of the same data segments extracted from the business data to be reviewed received by the target database.

[0018] Furthermore, under the condition that there is tampering in the process of allocating the business data to be audited received by the target database to the target node, the frequency of extracting the characteristic data of the business data to be audited is determined to be increased based on the comparison result that the data structure change value of the data segment extracted from the business data to be audited received by the supervision end is less than or equal to the data structure change value of the same data segment extracted from the business data to be audited received by the target database, and the frequency of extracting the characteristic data of the business data to be audited is determined to be increased based on the comparison result that the data structure change value of the data segment extracted from the business data to be audited received by the supervision end is greater than the data structure change value of the same data segment extracted from the business data to be audited received by the target database.

[0019] Furthermore, under the condition that all data is extracted from the business data to be reviewed received by the regulatory end and the business data to be reviewed received by the target database, it is determined that the process of allocating the business data to be reviewed to the target node has been tampered with based on the comparison result that the data volume of all data extracted from the business data to be reviewed received by the regulatory end is inconsistent with the data volume of all data extracted from the business data to be reviewed received by the target database.

[0020] Furthermore, under the condition that there is tampering in the process of allocating the business data to be audited received by the target database to the target node, based on the comparison result that the data property change value of all data extracted from the business data to be audited received by the supervision end is less than or equal to the data property change value of all data extracted from the business data to be audited received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be audited by a preset extraction frequency adjustment coefficient.

[0021] Furthermore, under the condition that there is tampering in the process of allocating the business data to be reviewed received by the target database to the target node, based on the comparison result that the data property change value of all data extracted from the business data to be reviewed received by the supervision end is greater than the data property change value of all data extracted from the business data to be reviewed received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be reviewed by a preset extraction frequency adjustment coefficient.

[0022] Compared with the prior art, the beneficial effect of the present invention is that the present invention generates business feature data by extracting feature data by different extraction methods, and generates a disguise mask by combining it with the node feature data of the target database node, and sends the business data to be reviewed and the disguise mask to the supervision end. After the supervision end completes data reception, the business data to be reviewed is allocated to the target database node for processing. By generating a disguise mask and sending it together with the business data to be reviewed, the leakage or tampering of data during the transmission process is effectively prevented, the security of data transmission is enhanced, and the business data to be reviewed and the disguise mask are sent to the supervision end, so that the supervision end can fully and accurately understand the true situation of the data, which provides strong support for subsequent audit work, thereby improving the efficiency of financial business audit processing.

[0023] Furthermore, the present invention determines the comparison method of the business data to be reviewed between the regulatory end and the target database by comparing the comprehensive evaluation value of the business data to be reviewed with the preset comprehensive evaluation value, thereby improving the flexibility and efficiency of data processing and improving the accuracy of financial business review processing.

[0024] Furthermore, the present invention determines whether the data volume of the business data to be audited is consistent between the regulatory end and the target database under the first comparison method by comparing the data volume of the same segment data in the regulatory end and the target database, and then infers whether there is any tampering in the data distribution process. It effectively utilizes the results of data segment comparison to quickly detect whether there is any tampering in the data during transmission or distribution, improves the verification efficiency of data integrity and security, provides a strong guarantee for the reliability of data processing, improves the accuracy of financial business audit processing, and thus improves the efficiency of financial business audit processing.

[0025] Furthermore, the present invention determines the extraction frequency of the feature data extraction method that is dynamically adjusted by comparing the ratio of the change values ​​of the same data structure in the supervision end and the target database with the preset ratio to cope with the risk of data tampering, thereby enhancing the system's responsiveness and flexibility to potential data tampering, helping to more accurately identify and respond to data security issues, and improving the security of overall data processing.

[0026] Furthermore, the present invention compares the data volume of all data in the regulatory end and the target database under the second comparison method, determines whether the data volume of the business data to be audited is consistent between the two, and then infers whether there is any tampering in the data distribution process, thereby improving the accuracy and reliability of data integrity verification, helping to quickly identify and respond to data tampering risks, ensuring the authenticity and security of data processing, and improving the accuracy of financial business audits, thereby improving the efficiency of financial business audit processing.

[0027] Furthermore, the present invention adjusts the extraction frequency of feature data by comparing the data property change values ​​of business data in the regulatory end and the target database. Through refined data property change value analysis, intelligent adjustment of the feature data extraction frequency is achieved, thereby improving the system's sensitivity and response capabilities to tampering risks, thereby improving the efficiency of financial business audit processing.

[0028] Furthermore, the present invention uses historical pending business data and audit results as training data to construct and train a large business model, and then uses the model to perform automated audits on subsequent pending businesses. By building a large model, the audit efficiency and accuracy are improved. As the amount of data increases, the model can be continuously optimized to adapt to new business scenarios and changes, thereby improving the efficiency of financial business audit processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of a financial business review processing method based on a large model according to an embodiment of the present invention;

[0030] Figure 2 A flowchart of a method for extracting characteristic data of business data to be reviewed according to an embodiment of the present invention;

[0031] Figure 3 A flowchart of an embodiment of the present invention for determining a method for comparing business data to be reviewed received by a supervisory terminal with business data to be reviewed received by a target database;

[0032] Figure 4 The present invention is a flowchart of a method for extracting characteristic data of business data to be reviewed according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0034] It should be pointed out that the data in this embodiment are obtained by comprehensive analysis and evaluation of the historical test data and the corresponding historical test results of the present invention in the three months before this test. It can be understood by those skilled in the art that the present invention can determine the above parameters for a single item by selecting the highest proportion of values ​​as the preset standard parameters according to the data distribution, using weighted summation to use the obtained values ​​as the preset standard parameters, substituting each historical data into a specific formula and using the values ​​obtained by the formula as the preset standard parameters or other selection methods, as long as the present invention can clearly define different specific situations in the single determination process through the obtained values.

[0035] See also Figure 1-Figure 4 As shown, Figure 1 It is a flow chart of a financial business review processing method based on a large model according to an embodiment of the present invention; Figure 2 A flowchart of a method for extracting characteristic data of business data to be reviewed according to an embodiment of the present invention; Figure 3 A flowchart of an embodiment of the present invention for determining a method for comparing business data to be reviewed received by a supervisory terminal with business data to be reviewed received by a target database; Figure 4 The present invention is a flowchart of a method for extracting characteristic data of business data to be reviewed according to an embodiment of the present invention.

[0036] The financial business review processing method based on the big model in the embodiment of the present invention includes:

[0037] Step S1, obtaining the to-be-audited business data sent by several sending ends, and analyzing the to-be-audited business data of each sending end to extract business characteristic data of the to-be-audited business data;

[0038] Step S2, obtaining node feature data of several nodes in the target database;

[0039] Step S3, randomly combining the service feature data and the node feature data to generate a camouflage mask;

[0040] Step S4, sending the to-be-audited business data and the disguised mask to a supervisory terminal;

[0041] Step S5, when the sending end completes sending the business data to be reviewed, the supervisory end distributes the business data to be reviewed received by the target database to the target node and determines whether the distribution process of the business data to be reviewed has been tampered with;

[0042] Step S6, using the business data to be reviewed and the review results in the historical supervision process of the supervision end on the business data to be reviewed as a review data set to train the business big model, and reviewing the subsequent business to be reviewed based on the trained big model.

[0043] Specifically, the embodiment of the present invention determines the analysis method of the characteristic data of the business data to be reviewed based on the comparison result of the importance evaluation value of the business data to be reviewed and the preset importance evaluation value of 0.85 under the condition of obtaining the business data to be reviewed sent by several sending ends;

[0044] When the importance evaluation value is less than or equal to the preset importance evaluation value, it is determined to extract the characteristic data of the to-be-audited business data in a first analysis manner;

[0045] When the importance evaluation value is greater than the preset importance evaluation value, it is determined to extract the characteristic data of the to-be-audited business data in a second analysis manner;

[0046] Among them, the business data to be reviewed includes but is not limited to business application data, transaction data and market data.

[0047] In the embodiment of the present invention, the first analysis method is to extract characteristic data of the business data to be reviewed based on the data structure change value of the business data to be reviewed, and the second analysis method is to extract characteristic data of the business data to be reviewed based on the data property change value of the business data to be reviewed.

[0048] In an embodiment of the present invention, the data structure change value is the increase or decrease in the data segment length of the business data to be reviewed, the data property change value is the decrease in the data importance of the business data to be reviewed, and the characteristic data is the changed business data to be reviewed.

[0049] In the embodiment of the present invention, the preset importance evaluation value is 0.85, but the above value is not limited thereto, and those skilled in the art may also adjust the value according to actual needs.

[0050] Specifically, the importance evaluation value is the difference between the number of customers and the historical number of customers divided by the historical number of customers multiplied by the ratio of the financial business processing time to the average historical financial business processing time.

[0051] In the embodiment of the present invention, the customer volume refers to the number of customers who apply for or participate in financial services, and the financial service processing time refers to the time required for a customer to complete the financial service processing from submitting an application for the financial service.

[0052] Specifically, the present invention determines the analysis method of characteristic data of the business data to be audited by comparing the importance evaluation value of the business data to be audited with the preset importance evaluation value, thereby improving the efficiency of data processing and the accuracy of the audit results, thereby improving the efficiency of financial audit processing.

[0053] Specifically, in the embodiment of the present invention, the feature data extracted by the first analysis method or the feature data extracted by the second analysis method is named as business feature data.

[0054] Specifically, in the embodiment of the present invention, the data characteristics of the target database node are named as node characteristic data, and the service characteristic data and the node characteristic data are randomly combined to generate a camouflage mask.

[0055] Specifically, the embodiment of the present invention sends the to-be-audited business data and the disguise mask to the supervision end.

[0056] Specifically, when the sending end completes sending the pending business data, the supervisory end distributes the pending business data received by the target database to the target node.

[0057] Specifically, the present invention generates business feature data by extracting feature data by different extraction methods, and generates a disguise mask by combining it with the node feature data of the target database node, and sends the business data to be reviewed and the disguise mask to the supervision end. After the supervision end completes data reception, the business data to be reviewed is distributed to the target database node for processing. By generating a disguise mask and sending it together with the business data to be reviewed, the data is effectively prevented from being leaked or tampered with during the transmission process, the security of data transmission is enhanced, and the business data to be reviewed and the disguise mask are sent to the supervision end, so that the supervision end can fully and accurately understand the true situation of the data, which provides strong support for subsequent audit work, thereby improving the efficiency of financial business audit processing.

[0058] Specifically, the embodiment of the present invention determines the method of comparing the business data to be reviewed received by the supervision end with the business data to be reviewed received by the target database according to the comparison result of the comprehensive evaluation value of the business data to be reviewed and the preset comprehensive evaluation value, under the condition that the business data to be reviewed received by the target database is allocated to the target node;

[0059] When the comprehensive evaluation value is less than or equal to the preset comprehensive evaluation value, determining a first comparison mode;

[0060] When the comprehensive evaluation value is greater than the preset comprehensive evaluation value, determining a second comparison method;

[0061] Among them, the first comparison method is to extract a piece of data with the same characteristics from the business data to be reviewed received by the regulatory end and the business data to be reviewed received by the target database; the second comparison method is to extract all data from the business data to be reviewed received by the regulatory end and the business data to be reviewed received by the target database, and the characteristics include changes in data structure and changes in data properties.

[0062] In the implementation of the present invention, the preset comprehensive evaluation value is 0.8, but the above value is not limited to this, and those skilled in the art can also adjust the value according to actual needs.

[0063] Specifically, the comprehensive evaluation value is the ratio of trading volume to historical trading volume multiplied by the ratio of trading times to the average historical trading times.

[0064] In the embodiment of the present invention, the transaction volume refers to the total amount, total quantity or total value involved in the financial business activities. The transaction times refers to the total number of transactions or business dealings occurring in the financial business activities.

[0065] Specifically, the present invention determines the comparison method of the business data to be reviewed between the regulatory end and the target database by comparing the comprehensive evaluation value of the business data to be reviewed with the preset comprehensive evaluation value, thereby improving the flexibility and efficiency of data processing and improving the accuracy of financial business audit processing.

[0066] Specifically, under the condition of determining the first comparison method, the embodiment of the present invention determines whether the data volume of the business data to be reviewed in the target database and the supervisory end is consistent according to the comparison result of the data volume of the data segment extracted from the business data to be reviewed received from the supervisory end and the data volume of the same data segment extracted from the business data to be reviewed received from the target database;

[0067] When the amount of data of the data segment extracted from the business data to be reviewed received by the supervisory end is consistent with the amount of data of the same data segment extracted from the business data to be reviewed received from the target database, it is determined that the amount of data of the business data to be reviewed in the target database and the supervisory end is the same;

[0068] When the data volume of the data segment extracted from the business data to be reviewed received by the supervisory end is inconsistent with the data volume of the same data segment extracted from the business data to be reviewed received from the target database, it is determined that the data volume of the business data to be reviewed in the target database and the supervisory end is different.

[0069] Specifically, in an embodiment of the present invention, based on the first comparison method, according to the comparison result that the data amounts of the business data to be audited in the target database and the regulatory end are the same, it is determined that there is no tampering in the process of allocating the business data to be audited received by the target database to the target node; based on the first comparison method, according to the comparison result that the data amounts of the business data to be audited in the target database and the regulatory end are different, it is determined that there is tampering in the process of allocating the business data to be audited received by the target database to the target node.

[0070] Specifically, the present invention compares the data volume of the same segment data in the regulatory end and the target database under the first comparison method, determines whether the data volume of the business data to be audited is consistent between the two, and then infers whether there is any tampering in the data distribution process. It effectively utilizes the results of data segment comparison to quickly detect whether there is any tampering in the data during transmission or distribution, improves the verification efficiency of data integrity and security, provides a strong guarantee for the reliability of data processing, improves the accuracy of financial business audit processing, and thus improves the efficiency of financial business audit processing.

[0071] Specifically, under the condition that it is determined that the process of allocating the business data to be reviewed received by the target database to the target node has been tampered with, the embodiment of the present invention determines to adjust the extraction method of the feature data of the business data to be reviewed according to the comparison result of the data structure change value of the data segment extracted from the business data to be reviewed received by the supervision end and the data structure change value of the same data segment extracted from the business data to be reviewed received by the target database;

[0072] When the data structure change value of the data segment extracted from the business data to be reviewed received by the supervision end is less than or equal to the data structure change value of the same data segment extracted from the business data to be reviewed received by the target database, it is determined to increase the extraction frequency of the feature data of the business data to be reviewed under the first analysis mode;

[0073] When the data structure change value of the data segment extracted from the business data to be reviewed received by the supervision end is greater than the data structure change value of the same data segment extracted from the business data to be reviewed received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be reviewed under the second analysis method.

[0074] Specifically, under the condition of determining to increase the extraction frequency of the first analysis method, the embodiment of the present invention determines to increase the extraction frequency of the feature data of the business data to be reviewed under the first analysis method according to the comparison result of the ratio of the data structure change value of the data segment extracted from the business data to be reviewed received by the supervision end and the data structure change value of the same data extracted from the business data to be reviewed received by the target database and a first preset ratio;

[0075] When the ratio is less than or equal to the first preset ratio, it is determined to increase the extraction frequency of the feature data of the to-be-audited business data under the first analysis mode to a corresponding value by using the first preset increase extraction adjustment coefficient of 1.03;

[0076] When the ratio is greater than the first preset ratio, it is determined to increase the extraction frequency of the feature data of the to-be-audited business data under the first analysis mode to a corresponding value by using the second preset increase extraction adjustment coefficient of 1.07;

[0077] The ratio is the ratio of the data structure change value of the data segment extracted from the pending business data received by the supervision end to the data structure change value of the same data segment extracted from the pending business data received by the target database.

[0078] In the embodiment of the present invention, the first preset ratio is set to 0.76, but the above value is not limited thereto, and those skilled in the art may also adjust the value according to actual needs.

[0079] In an embodiment of the present invention, the increased extraction frequency of the first analysis method is set to Mz, and Mz=M×Ki is set, wherein M represents the extraction frequency of the first analysis method, Ki represents the i-th preset increased extraction adjustment coefficient, i takes a value of 1 or 2, K1 is the first preset increased extraction adjustment coefficient, and K2 is the second preset increased extraction adjustment coefficient.

[0080] Specifically, under the condition of determining to increase the extraction frequency of the second analysis method, the embodiment of the present invention determines to increase the extraction frequency of the feature data of the business data to be reviewed under the second analysis method according to the comparison result of the ratio of the data structure change value of the data segment extracted from the business data to be reviewed received by the supervision end and the data structure change value of the same data segment extracted from the business data to be reviewed received by the target database and the second preset ratio;

[0081] When the ratio is less than or equal to the second preset ratio, it is determined that the extraction frequency of the characteristic data of the to-be-audited business data under the second analysis mode is increased to a corresponding value by using the first preset increase frequency adjustment coefficient of 1.06;

[0082] When the ratio is greater than the second preset ratio, it is determined that the extraction frequency of the feature data of the to-be-audited business data under the second analysis mode is increased to a corresponding value by using the second preset increase frequency adjustment coefficient of 1.13;

[0083] The ratio is the ratio of the data structure change value of the data segment extracted from the pending business data received by the supervision end to the data structure change value of the same data segment extracted from the pending business data received by the target database.

[0084] In the embodiment of the present invention, the second preset ratio is set to 0.73, but the above value is not limited thereto, and those skilled in the art may also adjust the value according to actual needs.

[0085] In an embodiment of the present invention, the extraction frequency of the increased second analysis method is set to Nz, and Nz=N×Tj is set, wherein N represents the extraction frequency of the second analysis method, Tj represents the jth preset increase frequency adjustment coefficient, j takes a value of 1 or 2, T1 is the first preset increase frequency adjustment coefficient, and T2 is the second preset increase frequency adjustment coefficient.

[0086] Specifically, the present invention determines the extraction frequency of the feature data extraction method that is dynamically adjusted by comparing the ratio of the change values ​​of the same data structure in the supervision end and the target database with the preset ratio to cope with the risk of data tampering, thereby enhancing the system's responsiveness and flexibility to potential data tampering, helping to more accurately identify and respond to data security issues, and improving the security of overall data processing.

[0087] Specifically, under the condition of determining the second comparison method, the embodiment of the present invention determines whether the data volume of the business data to be reviewed in the target database and the supervisory end is the same according to the comparison result of the data volume of all data extracted from the business data to be reviewed received from the supervisory end and the data volume of all data extracted from the business data to be reviewed received from the target database;

[0088] When the amount of all data extracted from the pending business data received by the supervisory end is the same as the amount of all data extracted from the pending business data received by the target database, it is determined that the amount of the pending business data in the target database and the supervisory end is the same;

[0089] When the amount of all data extracted from the pending business data received by the supervisory end is different from the amount of all data extracted from the pending business data received by the target database, it is determined that the amount of the pending business data in the target database and the supervisory end is different.

[0090] Specifically, in an embodiment of the present invention, based on the second comparison method, according to the comparison result that the data amounts of the business data to be audited in the target database and the regulatory end are the same, it is determined that there is no tampering in the process of allocating the business data to be audited received by the target database to the target node; based on the second comparison method, according to the comparison result that the data amounts of the business data to be audited in the target database and the regulatory end are inconsistent, it is determined that there is tampering in the process of allocating the business data to be audited received by the target database to the target node.

[0091] Specifically, the present invention compares the data volume of all data in the regulatory end and the target database under the second comparison method, determines whether the data volume of the business data to be audited is consistent between the two, and then infers whether there is any tampering in the data distribution process, thereby improving the accuracy and reliability of data integrity verification, helping to quickly identify and respond to data tampering risks, ensuring the authenticity and security of data processing, and improving the accuracy of financial business audits, thereby improving the efficiency of financial business audit processing.

[0092] Specifically, under the condition that it is determined that the process of allocating the business data to be reviewed received by the target database to the target node has been tampered with, the embodiment of the present invention determines to adjust the extraction method of the characteristic data of the business data to be reviewed according to the comparison result of the data property change value of all data extracted from the business data to be reviewed received from the supervision end and the data property change value of all data extracted from the business data to be reviewed received from the target database;

[0093] When the data property change value of all data extracted from the business data to be reviewed received by the supervision end is less than or equal to the data property change value of all data extracted from the business data to be reviewed received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be reviewed under the first analysis mode to a corresponding value by using a preset extraction frequency adjustment coefficient;

[0094] When the data property change value of all data extracted from the business data to be reviewed received by the supervision end is greater than the data property change value of all data extracted from the business data to be reviewed received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be reviewed under the second analysis mode to a corresponding value using a preset extraction frequency adjustment coefficient.

[0095] Specifically, the embodiment of the present invention calculates a preset extraction frequency adjustment coefficient according to the difference between the importance evaluation value and the preset importance evaluation value.

[0096] Specifically, the preset extraction frequency adjustment coefficient is the difference between the importance evaluation value and the preset importance evaluation value divided by the preset importance evaluation value.

[0097] Specifically, the present invention adjusts the extraction frequency of feature data by comparing the data property change values ​​of business data in the regulatory end and the target database. Through refined analysis of the data property change values, the intelligent adjustment of the extraction frequency of feature data is achieved, thereby improving the system's sensitivity and response capabilities to tampering risks, thereby improving the efficiency of financial business audit processing.

[0098] Specifically, the embodiment of the present invention uses the pending business data and audit results in the supervision process of the supervision end on several historical pending business data as an audit data set to train the business big model, and audits subsequent pending business based on the trained big model.

[0099] Specifically, the present invention uses historical pending business data and audit results as training data to build and train a large business model, and then uses the model to perform automated audits on subsequent pending businesses. By building a large model, the audit efficiency and accuracy are improved. As the amount of data increases, the model can be continuously optimized to adapt to new business scenarios and changes, thereby improving the efficiency of financial business audit processing.

[0100] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A financial business audit processing method based on a large model, characterized in that: include: Acquire the business data to be reviewed sent by several sending ends, and analyze the business data to be reviewed of each sending end to extract business characteristic data of the business data to be reviewed; Obtain node feature data of several nodes in the target database; Randomly combining the service characteristic data and the node characteristic data to generate a camouflage mask; Sending the pending business data and the disguised mask to a supervisory end; When the sending end completes sending the business data to be reviewed, the supervisory end determines, based on the comparison result that the comprehensive evaluation value of the business data to be reviewed is less than or equal to the preset comprehensive evaluation value, to extract data segments with the same characteristics as those in the business data to be reviewed received from the supervisory end, where the characteristics include changes in data structure and changes in data properties; Under the condition that the data segments with the same characteristics as those in the business data to be reviewed received by the target database are extracted from the business data to be reviewed received by the supervisory end, it is determined that the process of allocating the business data to be reviewed to the target node has been tampered with based on the inconsistent comparison result of the data volume of the data segment extracted from the business data to be reviewed received by the supervisory end and the data volume of the same data segment extracted from the business data to be reviewed received by the target database; The business data to be reviewed and the review results in the historical supervision process of the supervision end on a number of business data to be reviewed are used as a review data set to train the business big model, and the subsequent business data to be reviewed are reviewed based on the trained big model.

2. The method for financial business audit processing based on a large model according to claim 1 is characterized in that: Under the condition of obtaining business data to be reviewed sent by several sending ends, based on the comparison result that the importance evaluation value of the business data to be reviewed is less than or equal to the preset importance evaluation value, it is determined to extract the characteristic data of the business data to be reviewed based on the data structure change value of the business data to be reviewed, so as to generate business characteristic data.

3. The financial business audit processing method based on a large model according to claim 1 is characterized in that: Under the condition of obtaining business data to be reviewed sent by several sending ends, based on the comparison result that the importance evaluation value of the business data to be reviewed is greater than the preset importance evaluation value, it is determined to extract the characteristic data of the business data to be reviewed based on the data property change value of the business data to be reviewed, so as to generate business characteristic data.

4. The method for financial business audit processing based on a large model according to claim 2 or 3, characterized in that: Under the condition that the business characteristic data is generated, based on the comparison result that the comprehensive evaluation value of the business data to be reviewed is greater than the preset comprehensive evaluation value, it is determined to extract all data from the business data to be reviewed received from the supervision end and the business data to be reviewed received from the target database.

5. The method for financial business audit processing based on a large model according to claim 1 is characterized in that: Under the condition that there is tampering in the process of allocating the business data to be audited received by the target database to the target node, it is determined to increase the extraction frequency of the characteristic data of the business data to be audited based on the comparison result that the data structure change value of the data segment extracted from the business data to be audited received by the supervision end is less than or equal to the data structure change value of the same data segment extracted from the business data to be audited received by the target database, and it is determined to increase the extraction frequency of the characteristic data of the business data to be audited based on the comparison result that the data structure change value of the data segment extracted from the business data to be audited received by the supervision end is greater than the data structure change value of the same data segment extracted from the business data to be audited received by the target database.

6. The method for financial business audit processing based on a large model according to claim 4 is characterized in that: Under the condition that all data is extracted from the business data to be reviewed received by the regulatory end and the business data to be reviewed received by the target database, it is determined that the process of allocating the business data to be reviewed to the target node has been tampered with based on the comparison result that the data volume of all data extracted from the business data to be reviewed received by the regulatory end is inconsistent with the data volume of all data extracted from the business data to be reviewed received by the target database.

7. The method for financial business audit processing based on a large model according to claim 6 is characterized in that: Under the condition that there is tampering in the process of allocating the business data to be audited received by the target database to the target node, based on the comparison result that the data property change value of all data extracted from the business data to be audited received by the supervision end is less than or equal to the data property change value of all data extracted from the business data to be audited received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be audited by a preset extraction frequency adjustment coefficient.

8. The method for financial business audit processing based on a large model according to claim 6 is characterized in that: Under the condition that there is tampering in the process of allocating the business data to be audited received by the target database to the target node, based on the comparison result that the data property change value of all data extracted from the business data to be audited received by the supervision end is greater than the data property change value of all data extracted from the business data to be audited received by the target database, it is determined to increase the extraction frequency of the characteristic data of the business data to be audited by a preset extraction frequency adjustment coefficient.

Images