An intelligent examination and approval method, system, readable storage medium and computer device

Abstract

The embodiment of the application provides a kind of intelligent examination and approval method, system, readable storage medium and computer equipment, it is related to the field of credit risk control examination and approval in financial industry, which comprises: the examination and approval problem is decomposed into basic query problem and index analysis problem by base big model;Call base big language model, form basic query SQL according to the field contained in basic query problem, and basic query SQL includes specific query field;Call base big language model, form index query SQL according to the quantitative analysis index contained in index analysis problem, nest basic query SQL into index query SQL, obtain final query SQL;Call database to execute final query SQL, obtain transaction record detail data for basic query SQL, and obtain quantitative analysis index statistical result for the index query SQL;Different problems are assembled to corresponding SQL query statement for quantitative query, which can make up big model illusion and unstable answer.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of credit risk control and approval in the financial industry, and in particular to an intelligent approval method and system, a readable storage medium and computer equipment. BACKGROUND

[0002] When a financial institution conducts approval of a customer's financial transaction flow file, it usually manually interprets and analyzes the financial transaction flow file line by line, and in the process of line-by-line interpretation and analysis, some general auxiliary tools (such as Excel) or custom-developed special systems (flow analysis systems based on fixed procedures, SQL and scripts) are used to reduce workload and improve approval quality.

[0003] In the process of implementing the present application, the applicant found that at least the following problems exist in the prior art:

[0004] In actual practice, whether using general auxiliary tools (such as Excel) or custom-developed special systems, there will be some problems that are difficult to deal with, for example, the financial transaction flow file cannot be processed if the form is unknown, only a single search method can be provided, only fixed indicators can be provided, and a large amount of repetitive work will be generated for the approval personnel. SUMMARY

[0005] The embodiments of the present application provide an intelligent approval method, system, readable storage medium and computer equipment, which can solve the above technical problems in the prior art.

[0006] To achieve the above purpose, in a first aspect, the embodiments of the present application provide an intelligent approval method, comprising:

[0007] After extracting and saving the to-be-identified financial transaction flow file provided by the to-be-approved customer with credit needs in the database, receiving an approval question raised by an approval personnel when approving the credit needs of the to-be-approved customer, and decomposing the approval question into a basic query question for querying the basic information of the transaction flow and an index analysis question for querying the quantitative analysis index by using a base large language model;

[0008] Calling the base large language model, assembling a basic query SQL for querying the basic information of the transaction flow according to the fields contained in the basic query question, the basic query SQL including specific query fields;

[0009] Calling the base large language model, assembling an index query SQL for querying the quantitative analysis index of the transaction flow according to the quantitative analysis index contained in the index analysis question, and nesting the basic query SQL into the index query SQL to obtain a final query SQL;

[0010] The database is invoked to execute the final query SQL, transaction record detail data is obtained for the basic query SQL, quantitative analysis index statistical results are obtained for the index query SQL, and quantitative analysis results are formed based on the transaction record detail data and the quantitative analysis index statistical results.

[0011] In a second aspect, an embodiment of the present application provides an intelligent approval system, comprising:

[0012] An approval problem decomposition unit is configured to, after extracting and storing a to-be-identified financial transaction flow file provided by a to-be-approved client with credit needs in a database, receive an approval problem raised by an approval personnel when approving the credit needs of the to-be-approved client, and decompose the approval problem into a basic query problem for querying basic information of the transaction flow and an index analysis problem for querying quantitative analysis indexes by using a base large language model.

[0013] A query SQL assembly unit is configured to invoke the base large language model, assemble a basic query SQL for querying the basic information of the transaction flow according to fields contained in the basic query problem, the basic query SQL including specific query fields, and invoke the base large language model to assemble an index query SQL for querying the quantitative analysis indexes of the transaction flow according to quantitative analysis indexes contained in the index analysis problem, and nest the basic query SQL in the index query SQL to obtain a final query SQL.

[0014] A quantitative analysis unit is configured to invoke a database to execute the final query SQL, obtain transaction record detail data for the basic query SQL, obtain quantitative analysis index statistical results for the index query SQL, and form quantitative analysis results based on the transaction record detail data and the quantitative analysis index statistical results.

[0015] In a third aspect, an embodiment of the present application provides a computer readable storage medium storing one or more programs, which when executed by a computer device, cause the computer device to perform the intelligent approval method described above.

[0016] In a fourth aspect, an embodiment of the present application provides a computer device, comprising:

[0017] A processor and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the intelligent approval method described above.

[0018] The above technical solution has the following beneficial effects: it can identify the format of financial transaction records issued by any financial institution, and can intelligently query and statistically analyze the corresponding data in the transaction record based on the questions asked by the approver. At the same time, by breaking down complex approval issues into multiple simple questions and assembling the simple questions into corresponding SQL query statements for quantitative querying, compared with directly asking questions to a large model based on approval issues and financial transaction record data, it can make up for the shortcomings of large model illusion, unstable answers, and user question text length constraints. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of an intelligent approval method according to an embodiment of the present invention;

[0021] Figure 2 This is a structural diagram of an intelligent approval system according to an embodiment of the present invention;

[0022] Figure 3 This is a complete flowchart of the import and analysis of financial transaction log files according to an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] like Figure 1 As shown, in conjunction with embodiments of the present invention, an intelligent approval method is provided, comprising:

[0025] S101: After extracting and storing the financial transaction records provided by the client with credit needs in the database, the system receives approval questions raised by the approver when approving the client's credit needs. The system decomposes the approval questions into basic query questions that query the basic information of the transaction records and indicator analysis questions that query quantitative analysis indicators through the base language model.

[0026] S102: calling a base large language model, assembling a basic query SQL for querying basic information of the transaction flow according to fields contained in the basic query question, the basic query SQL including specific query fields;

[0027] S103: calling the base large language model, assembling an index query SQL for querying quantitative analysis indexes of the transaction flow according to quantitative analysis indexes contained in the index analysis question, nesting the basic query SQL into the index query SQL to obtain a final query SQL;

[0028] S104: calling a database to execute the final query SQL, obtaining transaction record detail data for the basic query SQL, obtaining quantitative analysis index statistical results for the index query SQL, and forming a quantitative analysis result based on the transaction record detail data and the quantitative analysis index statistical results.

[0029] In the credit approval process of a financial institution, analyzing the financial transaction flow file provided by a customer is one of the important steps of risk identification. However, line-by-line interpretation and analysis of the financial transaction flow file is a time-consuming and labor-intensive work that is prone to errors and highly dependent on personal experience. In order to reduce the amount of repetitive work as much as possible and improve the approval quality, a system is needed to automatically and intelligently complete the related work.

[0030] The financial transaction flow file provided by the client with credit needs to be extracted and saved in the database, and the problem of unknown table style cannot be identified is solved by extracting and saving the data of the financial transaction flow file to be identified.

[0031] The approval questions of the approval personnel are decomposed into different basic query questions and index analysis questions, and the corresponding basic query SQL and index query SQL are formed according to the fields of each question, so that different indexes can be queried according to different questions, and the diversity of index query is realized. The base large language model is used to query in the database to obtain quantitative analysis results. Decomposing the approval questions can make up for the shortcomings of large model illusion and unstable answers caused by directly inputting the approval questions into the base large language model. In addition, the workload of the approval personnel is reduced, and the efficiency and accuracy of manual approval are improved.

[0032] Preferably, the intelligent approval method further comprises S105: extracting and saving the financial transaction flow file provided by the client with credit needs to be identified in the database, comprising:

[0033] S1051: receiving an original financial transaction flow file provided by a client with credit needs to be approved, and converting the original financial transaction flow file into a financial transaction flow file to be identified in the original expression manner;

[0034] S1052: Based on the base large language model, the financial transaction log file to be identified is identified, the metadata of the original financial transaction log file is obtained, and the metadata is represented in the original expression mode. The metadata is used to generate a table building SQL statement for storing data in the to-be-identified financial transaction log file;

[0035] S1053: Based on the metadata, the data in the to-be-identified financial transaction log file is disassembled, and the table building SQL statement is used to write the disassembled data into the table in the database where the metadata is located.

[0036] The original financial transaction log file is converted into a to-be-identified financial transaction log file in the original expression mode, and the metadata is in the original expression mode in the original financial transaction log file. This facilitates the to-be-approved customer to quickly identify and correspond to the to-be-approved customer in the later period. Compared with the prior art in which the fields of the original financial transaction log files of different financial institutions are all converted into the same field, the reaction time of the to-be-approved customer is reduced, and the to-be-approved customer is more convenient.

[0037] The original financial transaction log files of different financial institutions can be automatically identified, and the financial inflow can be automatically imported into the database without the intervention of developers (without manual table building and data import). No traditional IT development process (such as design table building and manual table building, development of expansion interface, testing, and online data import) is needed. Only the original financial transaction log file needs to be uploaded to the intelligent approval system to complete the conversion of the original financial transaction log file into data in the database table. Especially when the amount of work for connecting the original financial transaction log file of unknown financial institution transaction log format is very small, almost zero, the work efficiency is greatly improved.

[0038] Preferably, S1051: receiving the original financial transaction log file provided by the to-be-approved customer with credit needs, converting the original financial transaction log file into a to-be-identified financial transaction log file in the original expression mode, comprising:

[0039] The PDF format original financial transaction log file provided by the to-be-approved customer with credit needs is received, the original financial transaction log file is decrypted using triple data symmetric encryption (Triple Data Encryption Standard, 3DES), the decrypted original financial transaction log file is de-noised, and the de-noising includes de-watermarking and line folding. Line folding refers to the behavior of automatically moving the part exceeding the set width (such as window boundary, text box limit or page width) to the next line when a line of text exceeds the set width.

[0040] Extracting the data in the original financial transaction file after denoising processing, saving the data in the original financial transaction file into a TXT format file in turn, and identifying the page symbol and / or page number in the original financial transaction file through a large language model and a special tool program, setting the corresponding page number and page symbol in the TXT format file according to the identified page symbol and / or page number, and supplementing the broken words and / or broken lines of the original financial transaction file to form complete words and / or complete line contents, obtaining the to-be-identified financial transaction file.

[0041] In one example, user A (such as an operation officer) responsible for importing the original financial transaction file uploads the PDF format original financial transaction file provided by the to-be-approved client through the interface of the intelligent approval system. The intelligent approval system reads the PDF format original financial transaction file through a PDF reading special tool, including: PDF format decryption, watermark removal, content reading, page division and page number identification, and broken word / broken line completion processing, and extraction into a TXT format file.

[0042] In the TXT format file, the original expression method in the original financial transaction file is used, which will reduce the reaction time of the to-be-approved client compared to the prior art in which the fields of the original financial transaction files of different financial institutions are all converted into the same field, and the to-be-approved client is more convenient.

[0043] Preferably, S1052: identifying the to-be-identified financial transaction file based on the base large language model, obtaining the metadata of the original financial transaction file, and representing the metadata in the original expression method, generating a table building SQL statement for storing the data in the to-be-identified financial transaction file based on the metadata, including:

[0044] Based on the base large language model, the first page content of the to-be-identified financial transaction file is identified, and the metadata is extracted, and the metadata is represented in the original expression method. The metadata at least includes: table header field (each field contained in the table header is separately regarded as a column), original table header and page symbol (regarded as a column), format category of amount (each field is separately regarded as a column), specific query field (each field is separately regarded as a column), to-be-approved client name, and financial institution issuing the original financial transaction file. Wherein, the specific query field is derived from the table header field, and the table header field one-to-one corresponds to the fields included in the original table header of the to-be-identified financial transaction file.

[0045] The format category of the amount refers to how the transaction amount field is composed in the original financial transaction log file, which is used for subsequent accurate determination of income and expenditure amounts. There are three categories of amount formats:

[0046] Category C: Only has a "Transaction Amount" field, using positive and negative signs to represent income and expenses respectively to distinguish between income and expenses;

[0047] Category B: Represented by two fields, "Transaction Amount" and "Income / Expense Type". The amount is always a positive number, and income and expenditure are distinguished by the income / expense type.

[0048] Category A: Represented by two fields, "Expenditure Amount" and "Income Amount", with all amounts being positive numbers.

[0049] The header / page break refers to the original header row or page break in the original financial transaction log file. It is used to inform the subsequent data entry program how to distinguish between data and field names in the file, so as to ensure the universality of the transaction data entry processing program. The page break is used to determine what a page break is, and then use a specified symbol as the page break to implement pagination processing. The original header is the original header in the transaction log file, containing information fields. In the metadata, all the contents of the original header are treated as a single field.

[0050] Specific query field names refer to the set of DB (database) field names that need to be used when performing feature analysis queries on transaction records. These can include transaction purpose (salary, pawning, etc.), counterparty, transaction summary, remarks, and notes.

[0051] Generate a table creation SQL statement to construct and store the data in the financial transaction log file to be identified. The table creation SQL statement carries all the fields included in the metadata, that is, the fields included in the table creation SQL statement correspond one-to-one with the metadata.

[0052] Based on the metadata obtained above, we first determine whether it is a new style financial transaction record file. This is done by simultaneously judging three conditions: the name of the financial institution, the original table header, and the SQL statement used to create the table. If the metadata involved in the three conditions is exactly the same as the metadata in the existing table, or if the metadata involved in the three conditions does not match the metadata in the existing table, it is considered a new style. If it is a new style, we first create a database table based on all the fields carried in the SQL statement used to create the table, and then import the data in the financial transaction record file to be identified using the original expression method. Otherwise, we directly import the data in the financial transaction record file to be identified into the existing table in the import database.

[0053] Preferably, in S101, the approval question is decomposed into a basic query question for querying the basic information of the transaction flow and an index analysis question for querying the quantitative analysis index by the base large language model, including:

[0054] The base large language model is called to perform word segmentation on the approval question by the base large language model, extract a specific keyword representing a specific query field, and extract a table header field related to the specific keyword as a field contained in the basic query question. The field contained in the basic query question at least includes: the specific keyword, the transaction time range and the amount range.

[0055] The keyword of the quantitative analysis index is extracted to form the index analysis question for querying the quantitative analysis index, wherein the keyword of the quantitative analysis index at least includes the following fields: transaction counterpart, transaction month, transaction total amount, transaction number of accumulations, transaction average amount and amount standard deviation.

[0056] The user B (such as an approval personnel) responsible for analysis can ask questions to the intelligent approval system (intelligent agent) (or generated in advance by the intelligent approval system), and the base large language model such as LLM (Large Language Model) analyzes and decomposes the questions of the approval personnel to obtain the basic query question and the index analysis question, especially the table header field related to the specific keyword. The approval question is disassembled, which can make up for the large model illusion and unstable answers caused by directly inputting the approval question into the base large language model.

[0057] The basic query question and the index analysis question perform quantitative analysis on the original financial transaction flow file as needed. Four types of quantitative analysis dimensions are included: transaction amount summary, transaction frequency, amount fluctuation range and transaction time period. Based on these dimensions and in cooperation with the specific keyword (specific to the detailed specific query field), the financial transaction flow file is queried and sorted, so that from thousands of financial transaction flow files, the transaction flow record meeting the specified specific query field can be quickly located, such as transaction amount being too high, or transaction being too frequent, or transaction cycle being stable, or transaction record occurring in a specific time period.

[0058] Preferably, S102: calling the base large language model, assembling the basic query SQL for querying the basic information of the transaction flow according to the fields contained in the basic query question, the basic query SQL including the specific query field, including:

[0059] The base large language model is called to assemble the corresponding query condition according to the specific query field corresponding to the specific keyword in the basic query question by the base large language model, obtain the basic query SQL, and save the basic query SQL to the database.

[0060] The base large language model is called, the quantitative analysis index set contained in the problem is analyzed according to the index, the index query SQL for querying the quantitative analysis index of the transaction flow is assembled, the index query SQL is composed of different indexes such as index 1, index 2, etc., the basic query SQL is nested in the index query SQL, and the final query SQL is obtained; the integration of the basic query SQL and the statistical query SQL is realized based on the SQL nesting mode based on the characteristics that the second SQL includes the first SQL is the SQL itself.

[0061] Preferably, S103: calling a database, executing the final query SQL, obtaining transaction record detail data for the basic query SQL, and obtaining quantitative analysis index statistical results for the index query SQL, the quantitative analysis index statistical results including four-dimensional data: transaction amount summary, transaction frequency, amount fluctuation range, and transaction time period, and being returned in a JSON collection format, wherein KEY is a field name, and VALUE is a specific value.

[0062] Intelligent quantitative analysis can automatically perform quantitative analysis on the original financial transaction flow file provided by the to-be-approved client according to the demand specified index, for example, from the aspects of amount size, frequency, fluctuation range, and time range, and the like, so that the approval personnel can quickly analyze positive information elements and negative information elements from the transaction flow, thereby greatly reducing the workload of the approval personnel in checking and analyzing thousands of transaction flows one by one.

[0063] Preferably, the intelligent approval method further comprises:

[0064] S106: calling a base large language model, and analyzing based on the transaction record detail data and the quantitative analysis index statistical results to obtain qualitative analysis results of the to-be-identified financial transaction flow file;

[0065] Specifically, step S106: calling a base large language model, and analyzing based on the transaction record detail data and the quantitative analysis index statistical results to obtain qualitative analysis results of the to-be-identified financial transaction flow file, comprising:

[0066] S1061: submitting the transaction record detail data and the quantitative analysis index statistical results to the base large language model, and performing qualitative analysis by the base large language model to obtain a comprehensive analysis text of the financial transaction of the to-be-approved client;

[0067] S1062: According to the set qualitative analysis indicators (items of risk control checks) and the risk control types (including income stability, bad spending, transaction behavior anomaly, and implicit liabilities) corresponding to the qualitative analysis indicators, the content matched with the approval problem is obtained from the knowledge base as a judgment rule and basis based on the vector query method, and the content matched with the approval problem includes one of the following: regulations and systems, typical cases, and credit product introductions (for example, the regulations stipulate that stable income is monthly income volatility rate exceeding S%), and the transaction record detail data are submitted to the base large language model. The risk control types of the qualitative analysis indicators in the approval problem are qualitatively analyzed by the base large language model to obtain the qualitative indicator analysis result, that is, the qualitative analysis result of each type of qualitative analysis indicator, including income stability, bad spending, transaction behavior anomaly, and implicit liabilities.

[0068] S1063: The comprehensive analysis text of the financial transaction of the to-be-approved client and the qualitative indicator analysis result are submitted to the base large language model, and the base large language model is analyzed to obtain the overall qualitative analysis result of the to-be-identified financial transaction flow file.

[0069] The qualitative analysis is intelligent auxiliary approval, which is essentially used for qualitative analysis of transaction flow, and gives corresponding risk prompts and related approval suggestions by combining the existing knowledge of the base large language model, local knowledge base data, database data, and external data, which can be used to assist the approval personnel to issue approval opinions and suggestions.

[0070] Preferably, the intelligent approval method can further include:

[0071] S107: Store the metadata as process trace data in the database for subsequent manual correction of the quantitative analysis result and the qualitative analysis result.

[0072] S108: In the process of converting the original financial transaction flow file into the to-be-identified financial transaction flow file in the original expression manner, each transaction flow is associated with the page number of the original financial transaction flow file and saved in the database, which is used for directly locating the corresponding flow transaction in the original financial transaction flow file when the approval personnel reviews and traces the quantitative analysis result; so that the transaction flow can be easily located to the specific page of the original financial transaction flow file on the application end UI (User Interface, user interface) in the future.

[0073] S109: Save the basic query SQL in the database, which is directly used by the approval personnel when reviewing and tracing the transaction record detail data, and is input into the base large language model to output the corresponding transaction record detail data. The SQL statement returned by the detail query is applied to the application, and the application end can query the database to obtain transaction detail record data and directly show it to the user.

[0074] S1010: Save the index query SQL in the database, which is directly used by the approval personnel when reviewing and tracing the quantitative analysis index statistical results, and is input into the base large language model to output the corresponding quantitative analysis index statistical results. The SQL statement returned by the detail query is applied to the application, and the application end can query the database to obtain specific transaction detail record data (details that meet the current index characteristics), and directly show it to user B. The specific detail record data obtained through the above search corresponds to the index details, and the page number, transaction counterparty, and transaction time can be used to locate the specific rows of the table in the original PDF file. User B can check each piece of original data in the details.

[0075] The basic query SQL and the corresponding transaction record detail data, the index query SQL statement and the corresponding quantitative analysis index statistical results are returned to the business application end and saved in the database for subsequent traceability processing. This avoids repeated operations and repeated decomposition problems when user B queries in the future, improves the operation experience, and reduces the consumption of computing power.

[0076] In summary, the complete flowchart of the import of the original financial transaction flow file and intelligent approval is as shown in Figure 3

[0077] Import the original financial transaction flow file; read the original financial transaction flow file to form a to-be-recognized financial transaction flow file; extract metadata; store the metadata in the database; and decompose the data of the to-be-recognized financial transaction flow file and write it to the database.

[0078] The approval personnel proposes an approval question; the approval question is decomposed; the metadata of the database is read to assemble a basic query SQL; an index query SQL is assembled, which may include index 1 and index 2, etc.; the database is called to form quantitative analysis results, which are written to the database and shown to the approval personnel; and the database is called to form qualitative analysis results and show them to the approval personnel.

[0079] As shown in Figure 2 , an intelligent approval system is provided according to the embodiments of the present application, which comprises:

[0080] ​The approval problem decomposition unit 21 is configured to, after extracting and storing the to-be-identified financial transaction flow file provided by the to-be-approved client with credit needs in the database, receive an approval question raised by an approval personnel when the approval personnel approves the credit needs of the to-be-approved client, and decompose the approval question into a basic query question for querying the basic information of the transaction flow and an index analysis question for querying the quantitative analysis index by using the base large language model.

[0081] The query SQL assembly unit 22 is configured to call the base large language model, assemble a basic query SQL for querying the basic information of the transaction flow according to the fields contained in the basic query question, and the basic query SQL includes specific query fields.

[0082] The base large language model is called to assemble an index query SQL for querying the quantitative analysis index of the transaction flow according to the quantitative analysis index contained in the index analysis question, and the basic query SQL is nested in the index query SQL to obtain a final query SQL.

[0083] The quantitative analysis unit 23 is configured to call the database to execute the final query SQL, obtain transaction record detail data according to the basic query SQL, obtain quantitative analysis index statistical results according to the index query SQL, and form a quantitative analysis result based on the transaction record detail data and the quantitative analysis index statistical results.

[0084] In the credit approval process of a financial institution, analyzing the financial transaction flow file provided by a client is one of the important steps of risk identification. However, line-by-line interpretation and analysis of the financial transaction flow file is a time-consuming and labor-intensive work that is prone to errors and highly dependent on personal experience. In order to reduce the amount of repetitive work as much as possible and improve the approval quality, the system needs to automatically and intelligently complete the related work.

[0085] The to-be-identified financial transaction flow file provided by the to-be-approved client with credit needs is extracted and stored in the database, and the problem of unknown table patterns not being able to be identified is solved by extracting and storing the data of the to-be-identified financial transaction flow file.

[0086] The approval question of the approval personnel is decomposed into different basic query questions and index analysis questions, and the basic query SQL and the index query SQL corresponding to each question are formed according to the respective fields, so that different indexes can be queried according to different questions, and the diversity of index query is realized. The base large language model is used to query in the database to obtain the quantitative analysis result. The workload of the approval personnel is reduced, and the efficiency and accuracy of manual approval are improved.

[0087] Preferably, the intelligent approval system further comprises:

[0088] a data extraction unit configured to extract and store a to-be-identified financial transaction log file provided by a to-be-approved client with a credit demand in a database;

[0089] the data extraction unit comprises:

[0090] a data format conversion subunit configured to receive an original financial transaction log file provided by the to-be-approved client with the credit demand, and convert the original financial transaction log file into a to-be-identified financial transaction log file in an original expression manner;

[0091] a metadata extraction subunit configured to identify the to-be-identified financial transaction log file based on a base large language model, acquire metadata of the original financial transaction log file, represent the metadata in the original expression manner, and generate a table-building SQL statement for storing data in the to-be-identified financial transaction log file based on the metadata;

[0092] a data extraction and storage subunit configured to disassemble the data in the to-be-identified financial transaction log file based on the metadata, and use the table-building SQL statement to write the disassembled data into a table in the database where the metadata is located in the original expression manner.

[0093] Converting the original financial transaction log file into the to-be-identified financial transaction log file in the original expression manner and representing the metadata in the original expression manner in the original financial transaction log file facilitate the to-be-approved client to quickly identify and correspond to the to-be-approved client in a later check, and compared with the prior art in which fields of original financial transaction log files of different financial institutions are all converted into a same field, the reaction time of the to-be-approved client is reduced, and the to-be-approved client is more convenient.

[0094] The original financial transaction log files of different financial institutions can be automatically identified and automatically imported into the database, without the intervention of a developer (without manual table building and data importing), and without a traditional IT development process (for example, design table building and manual table building, development of an extension interface, testing, and online data importing), and only the original financial transaction log file needs to be uploaded to the intelligent approval system to complete the conversion of the original financial transaction log file into data in the database table. Especially, when the amount of work for connecting the original financial transaction log file of the unknown financial institution transaction log format is very small, almost zero, the work efficiency is greatly improved.

[0095] Preferably, the data format conversion subunit is specifically configured to:

[0096] The original financial transaction flow file in PDF format provided by the client with credit demand is received, the original financial transaction flow file is decrypted, and the decrypted original financial transaction flow file is denoised.

[0097] The data in the denoised original financial transaction flow file is extracted, and the data in the original financial transaction flow file is sequentially saved into a TXT format file, while the page separator and / or page number in the original financial transaction flow file are identified, the corresponding page number and / or page separator are set in the TXT format file according to the identified page separator and / or page number, and the broken words and / or broken lines in the original financial transaction flow file are supplemented to form complete words and / or complete line contents, thereby obtaining a financial transaction flow file to be identified.

[0098] The user A (for example, an operation officer) responsible for importing the original financial transaction flow file uploads the original financial transaction flow file in PDF format provided by the client to be approved through the interface of the intelligent approval system, and the intelligent approval system reads the original financial transaction flow file in PDF format through the PDF reading special tool developed by itself, including: PDF format decryption, watermark removal, content reading, page and page number identification, and broken word / broken line completion, etc. The extracted data is saved into a TXT format file.

[0099] In the TXT format file, the original expression method in the original financial transaction flow file is used, which will reduce the reaction time of the client to be approved compared with the prior art in which the fields of the original financial transaction flow files of various financial institutions are all converted into the same field, and the client to be approved is more convenient.

[0100] Preferably, the metadata extraction subunit is specifically used for:

[0101] The first page flow content of the financial transaction flow file to be identified is identified based on the base large language model, and the metadata is extracted and represented in the original expression method. The metadata at least includes the following: table header field (each field contained in the table header is individually as a column), original table header and page separator (as a column), amount format category (each field is individually as a column), specific query field (each field is individually as a column), client to be approved name, and financial institution issuing the original financial transaction flow file. The specific query field is derived from the table header field, and the table header field one-to-one corresponds to the fields included in the original table header of the financial transaction flow file to be identified.

[0102] The amount format category refers to how the transaction amount field in the original financial transaction flow file is composed, which is used for accurate judgment of the income and expenditure amount subsequently. The amount format category includes three categories:

[0103] Class C: only "transaction amount" field, positive and negative signs represent income and expenditure to distinguish between income and expenditure;

[0104] Class B: use "transaction amount", "type of income and expenditure" two fields to represent, its amount is positive, distinguished by type of income and expenditure;

[0105] Class A: "expenditure amount", "income amount" two fields to represent, its amount is positive.

[0106] Table header / page separator, refers to the original financial transaction flow file in the original table header row or page separator, used to tell the subsequent data into the program how to distinguish between file data and field name, to ensure the universality of flow data into the processing program; page separator for determining what is a page separator, then use the specified symbol as a page separator to achieve paging processing; wherein, the original table header is the transaction flow file in the original table header, contains the information field, in the metadata will be "original table header" in all the content as a field processing.

[0107] Specific query field name refers to the subsequent transaction flow for feature analysis query, the need to use the DB (database) field name set, such as transaction purposes (salary, pawn, etc.), but also can be transaction counterpart, transaction summary, addenda and notes.

[0108] Generate build storage data in the said to be identified financial transaction flow file table SQL statement, in the table SQL statement carrying the metadata includes all the fields, that is, the table SQL statement contains the field, one by one with the metadata.

[0109] According to the above obtained metadata, first judge whether it is a new style of financial transaction flow file, that is, according to the financial institution name, the original table header, and the table SQL statement three conditions are judged at the same time, when the three conditions related to the metadata and the metadata in the existing table is exactly the same, when the three conditions related to the metadata and the metadata in the existing table is not consistent, it is new style, if it is new style, first according to the table SQL statement carrying all the fields to create a database table, then the data in the to be identified financial transaction flow file using the original expression way into, otherwise the data in the to be identified financial transaction flow file directly into the existing table in the database.

[0110] Preferably, the approval problem decomposition unit 21 comprises:

[0111] The first decomposition subunit is configured to call a base large language model, perform word segmentation on the approval question by using the base large language model, extract a specific keyword representing a specific query field, and extract a table header field related to the specific keyword as a field included in a basic query question. The field included in the basic query question at least includes the specific keyword, a transaction time range, and an amount range.

[0112] The second decomposition subunit is configured to extract a keyword of a quantitative analysis index, and form an index analysis question for querying the quantitative analysis index. The keyword of the quantitative analysis index at least includes the following fields: a transaction counterparty, a transaction month, a transaction total amount, a cumulative number of transaction pens, a transaction average amount, and an amount standard deviation.

[0113] A user B responsible for analysis (such as an approval personnel) can ask questions to an intelligent approval system (an intelligent agent) or generate the questions in advance by the intelligent approval system. The base large language model such as an LLM large model analyzes and decomposes the questions of the approval personnel to obtain a basic query question and an index analysis question. In particular, a table header field related to the specific keyword is extracted. The approval question is decomposed, which can compensate for the large model illusion and unstable answers caused by directly inputting the approval question into the base large language model.

[0114] The basic query question and the index analysis question perform quantitative analysis on the original financial transaction flow file as needed. The quantitative analysis includes four dimensions: transaction amount summary, transaction frequency, amount fluctuation range, and transaction time period. Based on these dimensions and in cooperation with the specific keyword (specific to the detailed specific query field), the financial transaction flow file is queried and sorted. Thus, from thousands of financial transaction flow files, a transaction flow record meeting a specified specific query field can be quickly located, such as a transaction amount that is too high, or a transaction that is too frequent, or a stable transaction period, or a transaction record occurring in a specific time period.

[0115] The query SQL assembly unit 22 is specifically configured to:

[0116] The base large language model is called to assemble a corresponding query condition according to a specific query field corresponding to the specific keyword in the basic query question, and a basic query SQL is obtained.

[0117] The base large language model is called, and a quantitative analysis index set contained in the problem is analyzed according to the index to assemble an index query SQL for querying the quantitative analysis index of the transaction flow: the index query SQL is composed of different indexes such as index 1, index 2, and the like, the basic query SQL is nested in the index query SQL, and a final query SQL is obtained; the integration of the basic query SQL and the statistical query SQL is implemented based on the SQL nesting mode based on the characteristics that the second SQL includes the first SQL is the SQL itself.

[0118] Preferably, S103: calling the database to execute the final query SQL, obtaining transaction record detail data for the basic query SQL, and obtaining quantitative analysis index statistical results for the index query SQL, the quantitative analysis index statistical results including four-dimensional data: transaction amount summary, transaction frequency, amount fluctuation range, and transaction time period, and being returned in a JSON set format, wherein KEY is a field name, and VALUE is a specific value. Intelligent quantitative analysis can automatically perform quantitative analysis auxiliary analysis on the original financial transaction flow file provided by the to-be-approved customer according to the demand specified index, for example, from the aspects of amount size, frequency level, fluctuation range, and time range, and the like, so that the approval personnel can quickly analyze positive information elements and negative information elements from the transaction flow, and the workload of the approval personnel in checking and verifying thousands of flows one by one is greatly reduced.

[0119] Preferably, the intelligent approval system further comprises:

[0120] The quantitative analysis unit is configured to call the base large language model, and analyze the transaction record detail data and the quantitative analysis index statistical results to obtain qualitative analysis results of the to-be-identified financial transaction flow file.

[0121] The quantitative analysis unit is specifically configured to:

[0122] The transaction record detail data and the quantitative analysis index statistical results are submitted to the base large language model, qualitative analysis is performed through the base large language model, and a comprehensive analysis text of the financial transaction of the to-be-approved customer is obtained.

[0123] According to the set qualitative analysis index (item of risk control check) and the risk control type (for example, income stability, bad expenditure, etc.) corresponding to the qualitative analysis index, the content matched with the approval problem is obtained from the knowledge base based on the vector query method, and the content matched with the approval problem includes one of the following: rules and regulations, typical cases, and credit product introduction. The content matched with the approval problem and the transaction record detail data are submitted to the base large language model, the risk control type of the qualitative analysis index in the approval problem is qualitatively analyzed by the base large model, and the qualitative index analysis result is obtained, that is, the qualitative analysis result of each type of qualitative analysis index: for example, income stability or bad expenditure, etc.

[0124] The comprehensive analysis text of the financial transaction of the to-be-approved client and the qualitative index analysis result are submitted to the base large language model, and the base large language model is analyzed to obtain the qualitative analysis result of the whole to-be-identified financial transaction flow file.

[0125] Qualitative analysis is intelligent auxiliary approval, which is essentially used for qualitative analysis of transaction flow, combined with existing knowledge of the base large model, local knowledge base data, database data and external data, to give corresponding risk prompts and related approval suggestions, which can be used to assist approval personnel to issue approval opinions and suggestions.

[0126] Preferably, the intelligent approval method further comprises a traceability unit for storing metadata as process trace data into a database for subsequent manual correction of quantitative analysis results and qualitative analysis results.

[0127] In the process of converting the original financial transaction flow file into the to-be-identified financial transaction flow file using the original expression method, each transaction flow is associated with the page number of the original financial transaction flow file and saved in the database, which is used for directly locating the corresponding flow transaction in the original financial transaction flow file when the approval personnel reviews and traces the quantitative analysis result, so that the transaction flow can be easily located in the specific page of the original financial transaction flow file from the application end UI.

[0128] The basic query SQL is saved in the database, which is used for directly adopting the basic query SQL and inputting into the base large language model when the approval personnel reviews and traces the transaction record detail data, and outputting the corresponding transaction record detail data; the SQL language returned by the detail query is input into the application, and the application end can query the database to obtain the transaction detail record data and directly show it to the user.

[0129] The index query SQL is saved in the database, and when the approval personnel reviews and traces the quantitative analysis index statistical result, the index query SQL is directly used and input to the base large language model, and the corresponding quantitative analysis index statistical result is output. The SQL statement returned by the detail query is applied to the application end, and the application end can obtain specific transaction detail record data (details conforming to the current index characteristics) by querying the database, and directly shows the user B. The specific detail record data obtained by the above-mentioned searching corresponds to the index details, the page number, the transaction counterparty and the transaction time, so that the specific rows of the table in the original PDF file can be positioned, and the user B can check each piece of original data in the details.

[0130] The basic query SQL and the corresponding transaction record detail data, the index query SQL statement and the corresponding quantitative analysis index statistical result are returned to the business application end, and are saved in the database, for subsequent traceability processing. The subsequent user B query avoids repeated operation and repeated decomposition problem, the process of SQL statement generation, improves the operation experience, and reduces the consumption of computing power.

[0131] In combination with the embodiments of the present application, a computer readable storage medium is provided, which stores one or more programs, which, when executed by a computer device, cause the computer device to perform the intelligent approval method of any one of the foregoing.

[0132] In combination with the embodiments of the present application, a computer device is provided, which includes a processor and a memory arranged to store computer executable instructions, which, when executed, cause the processor to perform the intelligent approval method of any one of the foregoing.

[0133] The beneficial technical effects achieved by the embodiments of the present application are as follows:

[0134] 1. The original financial transaction flow file of any format is intelligently recognized by the large model, the table is intelligently built and the data is imported, and the technical problem of data extraction is solved.

[0135] 2. The financial transaction flow file of any financial institution can be intelligently recognized, and quantitative analysis and qualitative analysis can be automatically performed. On the technical level, the code does not need to be modified, the adaptation and development time of the flow file of each financial institution is saved (from 3-7 person-days to 0), and on the business level, the work efficiency and quality of the approval personnel are improved (from 2-3 hours to 10-20 minutes).

[0136] 3. Based on the knowledge base, the continuous collection, storage and analysis of the related business experience of the financial suggestion flow file analysis are supported, and the accumulation and inheritance of the business knowledge are realized.

[0137] 4. Based on the approval questions from the approvers, the system can break down the questions into SQL statements, compensating for shortcomings such as the illusion of a large model and unstable answers. By logging the SQL statements, the workload of review can be significantly reduced.

[0138] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process may be rearranged without departing from the scope of this disclosure. The appended method claims provide elements of various steps in an exemplary order and are not intended to limit the scope to the specific order or hierarchy described.

[0139] In the above detailed description, various features are combined together in a single embodiment to simplify this disclosure. This approach to disclosure should not be construed as reflecting an intention that embodiments of the claimed subject matter require more features than are explicitly stated in each claim. Rather, as reflected in the appended claims, the invention is presented with fewer features than all of the features of the single disclosed embodiment. Therefore, the appended claims are hereby explicitly incorporated into the detailed description, wherein each claim stands alone as a preferred embodiment of the invention.

[0140] The disclosed embodiments have been described above to enable any person skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments without departing from the spirit and scope of this disclosure. Therefore, this disclosure is not limited to the embodiments given herein, but is consistent with the broadest scope of the principles and novel features disclosed in this application.

[0141] The foregoing description includes examples of one or more embodiments. It is certainly impossible to describe all possible combinations of components or methods in order to describe the above embodiments, but those skilled in the art will recognize that further combinations and arrangements of the various embodiments are possible. Therefore, the embodiments described herein are intended to cover all such changes, modifications, and variations that fall within the scope of the appended claims. Furthermore, the term "comprising" as used in the specification or claims is interpreted in a manner similar to the term "including," as interpreted when used as a conjunction in the claims. Additionally, the use of any term "or" in the specification of the claims is intended to mean "non-exclusive or."

[0142] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, and steps described in connection with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans can implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present embodiments.

[0143] The various illustrative logical blocks, modules, and steps described in connection with the embodiments disclosed herein can be implemented or performed by a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor can be a microprocessor, but in the alternative, the general purpose processor can be any conventional processor, controller, microcontroller, or state machine. A processor can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

[0144] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is tangible. A storage medium can be coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor and the storage medium can reside as discrete components in a user terminal. The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present embodiments. The various methods described herein can be implemented on a computer readable medium to provide a computer program product. The methods described herein also can be practiced via computer- implemented steps executing on a computer system. The present disclosure is an enabled with the exception of the following clauses:

[0145] In one or more exemplary designs, the functions described can be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions can be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. Storage media can be any available media that can be accessed by a computer. By way of example, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. Also, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or other wire-based, fiber- optic based, or wireless based communications, then the coaxial cable, fiber optic cable, twisted pair, DSL, or other wire-based, fiber-optic based, or wireless based communications are included in the definition of medium. Disk and disc, as used herein, includes compact disc (CD), laser disc, optical disc, DVD, floppy disk, and Blu-ray® disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0146] The specific implementation described above is illustrative for purposes of teaching the present application. The application should not be limited, however, to these specific implementation. Any modifications, equivalents, or improvements, as long as they are within the spirit and scope of the present application are included in the scope of the present application.

Claims

1. An intelligent approval method, characterized in that, include: After extracting and storing the financial transaction records of the pending customers with credit needs in the database, the system receives approval questions raised by the approvers when approving the credit needs of the pending customers. The system decomposes the approval questions into basic query questions that query the basic information of the transaction records and indicator analysis questions that query quantitative analysis indicators through the base big language model. The base language model is invoked, and a basic query SQL is assembled based on the fields contained in the basic query question to query the basic information of the transaction flow. The basic query SQL includes specific query fields. The base language model is invoked, and quantitative analysis indicators included in the indicator analysis problem are assembled to form an indicator query SQL for querying the quantitative analysis indicators of transaction flow. The basic query SQL is nested into the indicator query SQL to obtain the final query SQL. The database is invoked to execute the final query SQL. Transaction record details are obtained from the basic query SQL, and quantitative analysis indicator statistics are obtained from the indicator query SQL. Quantitative analysis results are formed based on the transaction record details and the quantitative analysis indicator statistics. The transaction record details and the statistical results of the quantitative analysis indicators are submitted to the base big language model. Qualitative analysis is performed through the base big language model to obtain a comprehensive analysis text of the financial transactions of the client to be approved. Based on the set qualitative analysis indicators and the risk control types corresponding to the qualitative analysis indicators, the content matching the approval question is obtained from the knowledge base using the vector query method. The content matching the approval question includes one of the following: rules and regulations, typical cases, and credit product introductions. The content matching the approval question and the transaction record details are submitted to the base big language model. The base big language model performs qualitative analysis on the risk control types of the qualitative analysis indicators in the approval question to obtain the qualitative indicator analysis results. The comprehensive analysis text of the financial transactions of the client to be approved and the qualitative indicator analysis results are submitted to the base big language model. The base big language model is used for analysis to obtain the overall qualitative analysis results of the financial transaction flow file to be identified. The steps for extracting and storing the financial transaction records of pending clients with credit needs in the database include: Receive original financial transaction records from clients with credit needs awaiting approval, and convert the original financial transaction records into financial transaction records to be identified using the original wording. The financial transaction log file to be identified is identified based on the base big language model, the metadata of the original financial transaction log file is obtained, and the metadata is represented in the original expression method. Based on the metadata, SQL statements for building tables to store the data in the financial transaction log file to be identified are generated. The data in the financial transaction log file to be identified is decomposed based on the metadata, and the table creation SQL statement is used to write the decomposed data into the table in the database where the metadata is located, using the original expression method.

2. The intelligent approval method according to claim 1, characterized in that, The process of receiving original financial transaction records from clients with credit needs awaiting approval, and converting these original records into unidentified financial transaction records using their original format, includes: Receive original financial transaction records in PDF format from clients with credit needs awaiting approval, decrypt the original financial transaction records, and perform noise reduction processing on the decrypted original financial transaction records. Extract the data from the original financial transaction log file after noise reduction, and save the data from the original financial transaction log file sequentially into a TXT format file. At the same time, identify the page breaks and / or page numbers in the original financial transaction log file, set the corresponding page numbers and / or page breaks in the TXT format file according to the identified page breaks and / or page numbers, and supplement the words and / or lines in the original financial transaction log file to form complete words and / or complete lines of content, thus obtaining the financial transaction log file to be identified.

3. The intelligent approval method according to claim 1, characterized in that, The process involves identifying the financial transaction log file based on a base-based large language model, obtaining the metadata of the original financial transaction log file, representing the metadata using the original representation method, and generating SQL statements to construct tables and store the data within the financial transaction log file based on the metadata, including: The first page of the financial transaction record document to be identified is identified based on the base big language model, and metadata is extracted. The metadata is then represented using the original expression method. The metadata includes at least the following: header fields, original header and page break, amount format category, specific query fields, name of the customer to be approved, and the financial institution that issued the original financial transaction record document. The specific query fields are derived from the header fields, and the header fields correspond one-to-one with the fields included in the original header of the financial transaction record document to be identified. Generate a table creation SQL statement to construct and store the data in the financial transaction log file to be identified, and carry all fields included in the metadata in the table creation SQL statement.

4. The intelligent approval method according to claim 3, characterized in that, The approval problem is decomposed into a basic query problem (querying basic information about transaction flows) and an indicator analysis problem (querying quantitative analysis indicators) using a base-based large language model, including: The base language model is invoked to segment the approval question, extract specific keywords representing specific query fields, and extract table header fields related to the specific keywords as fields included in the basic query question. The fields included in the basic query question include at least: specific keywords, transaction time range, and amount range. Extract keywords from the quantitative analysis indicators to form an indicator analysis question for querying the quantitative analysis indicators. The keywords of the quantitative analysis indicators shall include at least the following fields: amount, cumulative number of transactions, average transaction amount, and standard deviation of amount.

5. The intelligent approval method according to claim 4, characterized in that, The invocation base language model assembles a basic query SQL based on the fields contained in the basic query question to query basic information about the transaction flow. The basic query SQL includes specific query fields, including: The base language model is invoked, and the corresponding query conditions are assembled based on the specific query fields corresponding to the specific keywords in the basic query question to obtain the basic query SQL. The basic query SQL is then saved to the database.

6. The intelligent approval method according to claim 1, characterized in that, Also includes: In the process of converting the original financial transaction log file into a financial transaction log file to be identified using the original expression method, each transaction log is associated with the page number of the original financial transaction log file and stored in the database. This is used so that when the approver reviews and traces the quantitative analysis results, the corresponding transaction log can be directly located in the original financial transaction log file. The basic query SQL is stored in the database and used by the approvers to directly use the basic query SQL and input it into the base language model when reviewing and tracing the transaction record details data, and output the corresponding transaction record details data. The indicator query SQL is stored in the database. When the approver reviews and traces the statistical results of the quantitative analysis indicators, they can directly use the indicator query SQL and input it into the base language model to output the corresponding quantitative analysis indicator statistical results.

7. An intelligent approval system, characterized in that, include: The approval issue decomposition unit is used to extract and save the financial transaction flow documents provided by the customers with credit needs to be identified in the database, and then receive the approval issues raised by the approvers when approving the credit needs of the customers. The approval issues are decomposed into basic query issues that query the basic information of the transaction flow and indicator analysis issues that query the quantitative analysis indicators through the base big language model. The query SQL assembly unit is used to call the base large language model and assemble a basic query SQL to query the basic information of the transaction flow based on the fields contained in the basic query question. The basic query SQL includes specific query fields. And call the base language model, assemble the quantitative analysis indicators contained in the indicator analysis problem to form the indicator query SQL for querying the quantitative analysis indicators of transaction flow, and nest the basic query SQL into the indicator query SQL to obtain the final query SQL. The quantitative analysis unit is used to call the database, execute the final query SQL, obtain transaction record details for the basic query SQL, obtain quantitative analysis indicator statistical results for the indicator query SQL, and form quantitative analysis results based on the transaction record details and the quantitative analysis indicator statistical results. The data extraction unit is used to extract and store the financial transaction records provided by customers with credit needs who are awaiting approval in the database. The quantitative analysis unit is specifically used for: The transaction record details and the statistical results of the quantitative analysis indicators are submitted to the base big language model. Qualitative analysis is performed through the base big language model to obtain the comprehensive analysis text of the financial transactions of the client to be approved. Based on the set qualitative analysis indicators and the risk control types corresponding to the qualitative analysis indicators, the content matching the approval question is obtained from the knowledge base using the vector query method. The content matching the approval question includes one of the following: rules and regulations, typical cases, and credit product introductions. The content matching the approval question and the transaction record details are submitted to the base big language model. The base big language model performs qualitative analysis on the risk control types of the qualitative analysis indicators in the approval question to obtain the qualitative indicator analysis results. The comprehensive analysis text of the financial transactions of the client to be approved and the qualitative indicator analysis results are submitted to the base big language model. The base big language model is used for analysis to obtain the overall qualitative analysis results of the financial transaction flow file to be identified. The data extraction unit includes: The data format conversion subunit is used to receive the original financial transaction records provided by the customer with credit needs who is awaiting approval, and convert the original financial transaction records into a financial transaction record to be identified using the original expression method. The metadata extraction subunit is used to identify the financial transaction log file to be identified based on the base big language model, obtain the metadata of the original financial transaction log file, represent the metadata in the original expression method, and generate SQL statements for building tables to store the data in the financial transaction log file to be identified based on the metadata. The data extraction and storage subunit is used to decompose the data in the financial transaction log file to be identified based on the metadata. The table creation SQL statement is used to write the decomposed data into the table in the database where the metadata is located, using the original expression method.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which, when executed by a computer device, cause the computer device to perform the intelligent approval method according to any one of claims 1-6.

9. A computer device, characterized in that, include: processor; And a memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the intelligent approval method according to any one of claims 1-6.

Images