Small and micro credit approver expert system and method based on large language model

Abstract

The invention provides a small and micro credit approver expert system and method based on a large language model, and belongs to the technical field of approval management, and the system specifically comprises a problem point recording module which is responsible for recording problem points of a credit report in an approval processing process, and a potential problem point output module which is responsible for outputting potential problem points on the basis of the credit report and the problem points of the credit report. The method comprises the steps that a big language model is used for carrying out recognition processing to obtain potential problem points, a retrieval processing module carries out retrieval processing on the basis of the potential problem points to obtain supplementary retrieval information of the potential problem points, and a conclusion output module is responsible for outputting evaluation conclusions of credit loan reports according to the potential problem points and the corresponding supplementary retrieval information. And the approval processing efficiency of the credit report is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of approval management, and in particular relates to a small and micro enterprise credit approval officer expert system and method based on a large language model. Background Art

[0002] In the field of small and micro enterprise credit, the review of credit reports is highly difficult and involves a large workload. Facing thousands of industries, the customer situation is complex and changeable. The data obtained through expert judgment and the credit investigation process is limited, making it difficult to ensure the approval effect.

[0003] To achieve the approval processing of credit reports, for example, in the invention patent application CN202011435869.0 "A Method for Intelligent Analysis of Credit Customers", a credit customer model is established based on account settlement transaction flow data, and a credit report is generated according to the credit customer model, making the data in the credit report more unified, comprehensive, detailed, and accurate, and the credit report more standardized, thus facilitating the improvement of approval efficiency.

[0004] During the approval process of credit reports, the amount of data involved in credit reports is large. In existing technical solutions, the personal experience of approval officers is often relied on. Therefore, not only is the accuracy rate of identifying and processing problem items difficult to meet the requirements, but also data retrieval cannot be carried out specifically during the approval process to identify and process problem items, resulting in the inability to meet the requirements for both the accuracy and efficiency of the approval processing of credit reports.

[0005] To solve the above technical problems, the present application provides a small and micro enterprise credit approval officer expert system and method based on a large language model. Summary of the Invention

[0006] To achieve the purpose of the present invention, the present invention adopts the following technical solutions: In a first aspect, the present application provides a small and micro enterprise credit expert system based on a large language model, specifically including: A problem point recording module, a potential problem point output module, a retrieval processing module, and a conclusion output module; The problem point recording module is responsible for recording the problem points in the approval process of the credit report; The potential problem point output module is responsible for using the large language model to perform identification processing based on the credit report and the problem points of the credit report to obtain potential problem points; The retrieval processing module performs retrieval processing based on the potential problem points to obtain supplementary retrieval information for the potential problem points; The conclusion output module is responsible for outputting the evaluation conclusion of the credit report based on the potential problem points and the corresponding supplementary retrieval information.

[0007] The beneficial effects of the present invention are as follows: Based on the credit report and the problem points of the credit report, the large language model is used for identification and processing to obtain potential problem points, thus avoiding the technical problems that the efficiency and accuracy of the original review and processing relying solely on the personal experience of the examiner do not meet the requirements. Considering the advantages of the large language model in the data analysis and reasoning process, it also lays a foundation for further retrieval and analysis based on the potential problem points.

[0008] Based on the potential problem points and the corresponding supplementary retrieval information, the evaluation conclusion of the credit report is output, thus avoiding the technical problems that the reliability of the approval result does not meet the requirements due to solely considering the potential problem points and the data of the credit report. It improves the accuracy and reliability of the overall approval process of the credit report, and at the same time ensures the efficiency of the approval process of the credit report.

[0009] A further technical solution is that the problem points include credit data, credit usage data, production and operation indicators, and financial indicators.

[0010] A further technical solution is that the supplementary retrieval information is the retrieval result of data associated with the potential problem points.

[0011] A further technical solution is that the output of the evaluation conclusion of the credit report specifically includes: Taking the potential risk points and the supplementary retrieval information of the potential risk points as input quantities, and using the output result of the large model to determine the evaluation conclusion of the credit report.

[0012] In the second aspect, the present application provides a small and micro enterprise credit approval method based on a large language model, which is applied to the above-mentioned small and micro enterprise credit expert system based on a large language model, and specifically includes: S1 Record the problem points in the approval process of the credit report, determine the associated report content of different problem points according to the types of the problem points, and when it is determined that the risks of the credit report meet the requirements, proceed to the next step; S2 Based on the credit report and the problem points of the credit report, use the large language model for identification and processing to obtain potential problem points, and determine the target retrieval source according to the information association and update situation between different retrieval sources and the potential problem points; S3 Based on the retrieval processing results of the target retrieval source, determine the deviation situation of the supplementary retrieval information of different potential risk points. When it is determined that the retrieval plan needs to be adjusted based on the deviation situation, obtain the supplementary retrieval information of the credit report based on the full amount of retrieval sources; S4 Output the evaluation conclusion of the credit report based on the potential problem points and the corresponding supplementary retrieval information.

[0013] A further technical solution is that the problem points in the approval process are determined according to the problems in the approval process of relevant personnel in the credit report.

[0014] A further technical solution is that the associated report content of the problem points is determined according to the report content where the problem points exist.

[0015] A further technical solution is that determining that the risk of the credit report meets the requirements specifically includes: Based on the associated report content of the problem points, determine the associated report content in different sub-units in the credit report; According to the constituent data of the associated report content in different sub-units, determine the problem sub-units of the associated report content; Determine whether the risk of the credit report meets the requirements by the number of the problem sub-units.

[0016] A further technical solution is that when the risk of the credit report does not meet the requirements, then based on all the indicators and based on the full amount of retrieval sources, obtain the supplementary retrieval information of the credit report.

[0017] A further technical solution is that the sub-units include a credit analysis sub-unit, a production and operation sub-unit, and a financial analysis sub-unit.

[0018] A further technical solution is that determining that the retrieval plan needs to be adjusted specifically includes: Based on the deviation situation of the supplementary retrieval information of different potential risk points in the target data source, determine the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information; Based on the deviation situation between the associated report content and the supplementary retrieval information, determine the number of potential risk points of the associated report content with deviation; According to the number of potential risk points of the associated report content with deviation, determine whether the retrieval plan needs to be adjusted.

[0019] A further technical solution is that when the number of potential risk points of the associated report content with deviation is greater than the preset risk point number threshold, it is determined that the retrieval plan needs to be adjusted.

[0020] A further technical solution is that when the retrieval plan does not need to be adjusted, then using the potential risk points and the supplementary retrieval information of the potential risk points in the target data source as input quantities, determine the evaluation conclusion of the credit report by the output result of the large model.

[0021] A further technical solution lies in that the evaluation conclusions of the credit report include pass and fail.

[0022] Other features and advantages will be described in the subsequent specification. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification and the drawings.

[0023] To make the above objectives, features, and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, provides a detailed description as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] By referring to the accompanying drawings and describing its exemplary embodiments in detail, the above and other features and advantages of the present invention will become more obvious.

[0025] Figure 1 is a framework diagram of a small and micro enterprise credit expert system based on a large language model; Figure 2 is a flowchart of a small and micro enterprise credit approval method based on a large language model; Figure 3 is a flowchart for determining that the risks of the credit report meet the requirements; Figure 4 is a flowchart of a method for determining a target retrieval source; Figure 5 is a flowchart for determining that adjustment processing of the retrieval scheme needs to be performed. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] To enable those skilled in the art to better understand the technical solutions in this specification, the following will clearly and completely describe the technical solutions in the embodiments of this specification in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all the embodiments. Based on the embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this specification.

[0027] In this application, by using a large model to identify potential risk points of a credit report and further retrieving the potential risk points to obtain supplementary retrieval information, and using the supplementary retrieval information and the potential risk points to output the evaluation conclusion of the credit report based on the large model, the efficiency of the approval process of the credit report is improved.

[0028] Embodiment 1 As Figure 1 shown, this application provides a small and micro enterprise credit expert system based on a large language model, specifically including: A problem point recording module, a potential problem point output module, a retrieval processing module, and a conclusion output module; The problem point recording module is responsible for recording the problem points in the approval process of the credit report; The potential problem point output module is responsible for using the large language model to perform identification processing based on the credit report and the problem points of the credit report to obtain potential problem points; The retrieval processing module performs retrieval processing based on the potential problem points to obtain supplementary retrieval information for the potential problem points; The conclusion output module is responsible for outputting the evaluation conclusion of the credit report based on the potential problem points and the corresponding supplementary retrieval information.

[0029] Furthermore, the problem points include credit data, credit usage data, production and operation indicators, and financial indicators.

[0030] Specifically, the supplementary retrieval information is the retrieval result of data associated with the potential problem points.

[0031] It can be understood that the output of the evaluation conclusion of the credit report specifically includes: Using the potential risk points and the supplementary retrieval information of the potential risk points as input quantities, and determining the evaluation conclusion of the credit report based on the output result of the large model.

[0032] Embodiment 2 Second, as Figure 2 shown, the present application provides a small business credit approval method based on a large language model, which is applied to the above-mentioned small business credit expert system based on a large language model, and specifically includes: S1 Record the problem points in the approval process of the credit report, determine the associated report content of different problem points according to the types of the problem points, and when it is determined that the risks of the credit report meet the requirements based on the associated report content of different problem points, proceed to the next step; When the word count ratio of the associated report content of the problem points in the credit report is greater than 0.2, it is determined that the risks of the credit report do not meet the requirements.

[0033] S2 Use the large language model to perform identification processing based on the credit report and the problem points of the credit report to obtain potential problem points, and determine the target retrieval source according to the information association situation and update situation between different retrieval sources and the potential problem points; Based on the information association between the retrieval source and the potential problem points, determine the potential problem points of the retrieval source for supplementary retrieval information, and regard them as information association problem points. According to the data update cycle of the retrieval source, determine the timeliness coefficients of different information association problem points. Based on the sum of the timeliness coefficients of different information association problem points, determine the matching coefficient of the target retrieval source. The target retrieval source is the data source with the largest matching coefficient and all potential problem association points belonging to information association problem points.

[0034] S3 Based on the retrieval processing results of the target retrieval source, determine the deviation of the supplementary retrieval information of different potential risk points. When it is determined that the retrieval scheme needs to be adjusted based on the deviation, obtain the supplementary retrieval information of the credit report based on the full set of retrieval sources. Based on the deviation between the associated report content and the supplementary retrieval information, determine the number of potential risk points of the associated report content with deviation. When the number of potential risk points of the associated report content with deviation is greater than the preset risk point number threshold, it is determined that the retrieval scheme needs to be adjusted.

[0035] S4 Output the evaluation conclusion of the credit report based on the potential problem points and the corresponding supplementary retrieval information.

[0036] Furthermore, the problem points in the approval process are determined according to the problems in the approval process of the relevant personnel of the credit report.

[0037] It can be understood that the associated report content of the problem points is determined according to the report content with the problem points.

[0038] Specifically, as Figure 3 shown, determining that the risk of the credit report meets the requirements specifically includes: Based on the associated report content of the problem points, determine the associated report content in different sub-units of the credit report. According to the composition data of the associated report content in different sub-units, determine the problem sub-units of the associated report content. Determine whether the risk of the credit report meets the requirements by the number of the problem sub-units.

[0039] Furthermore, the problem sub-units are the sub-units where the word count ratio of the associated report content is greater than the preset word count ratio.

[0040] It should be noted that when the number of the problem sub-units is greater than the preset sub-unit number threshold, it is determined that the risk of the credit report does not meet the requirements.

[0041] It is understandable that when the risk of the credit report does not meet the requirements, based on all the indicators and on the basis of the full set of retrieval sources, the supplementary retrieval information of the credit report is obtained.

[0042] Further, the subunit includes a credit analysis subunit, a production and operation subunit, and a financial analysis subunit.

[0043] In another possible embodiment, determining that the risk of the credit report meets the requirements specifically includes: Based on the associated report content of the problem points, determining the word count ratio of the associated report content in the credit report; Based on the word count ratio of the associated report content, determining whether the risk of the credit report meets the requirements.

[0044] Further, when the word count ratio of the associated report content is not within the preset word count ratio range, it is determined that the risk of the credit report does not meet the requirements.

[0045] In another possible embodiment, determining that the risk of the credit report meets the requirements specifically includes: Obtaining the problem points of the credit report. When the number of problem points of the credit report does not meet the requirements, it is determined that the risk of the credit report does not meet the requirements; When the number of problem points of the credit report meets the requirements: Based on the number of problem points in different indicators, when it is determined that there is an indicator with the number of problem points not meeting the requirements, it is determined that the risk of the credit report does not meet the requirements; When there is no indicator with the number of problem points not meeting the requirements: Obtaining the indicators with problem points and combining the number of problem points in different indicators to determine the indicator risk coefficient of the credit report. When the indicator risk coefficient of the credit report is within the preset indicator risk coefficient range, it is determined that the risk of the credit report meets the requirements; When the indicator risk coefficient of the credit report is not within the preset indicator risk coefficient range: Based on the associated report content of the problem points, determining the associated report content in different subunits of the credit report. When the number of subunits with associated report content does not meet the requirements, it is determined that the risk of the credit report does not meet the requirements; When the number of subunits with associated report content meets the requirements, According to the composition data of the associated report content in different subunits, when it is determined that there is no problem subunit in the associated report content, it is determined that the risk of the credit report meets the requirements; When there are problem subunits in the associated report content: Obtain the number of problem subunits in the associated report content. When the number of problem subunits in the associated report content does not meet the requirements, it is determined that the risk of the credit report does not meet the requirements; When the number of problem subunits in the associated report content meets the requirements: Based on the constituent data of the associated report content in different subunits and in combination with the number of problem subunits, determine the report content risk coefficient of the credit report, and based on the report content risk coefficient, determine whether the risk of the credit report meets the requirements.

[0046] Further, when the report content risk coefficient does not meet the requirements, it is determined that the risk of the credit report does not meet the requirements.

[0047] Specifically, the indicators include credit data, credit usage data, production and operation indicators, and financial indicators.

[0048] Specifically, the potential problem points are determined according to the inference results of the large model.

[0049] It should be noted that as Figure 4 shown, the method for determining the target retrieval source is: Based on the information association situation between the retrieval source and the potential problem points, determine the potential problem points where the retrieval source has supplementary retrieval information, and use them as information association problem points; According to the data update cycle of the retrieval source, determine the timeliness coefficients of different information association problem points; Based on the sum of the timeliness coefficients of different information association problem points, determine the matching coefficient of the target retrieval source, and based on the matching coefficient, determine whether the data source is the target retrieval source.

[0050] Further, the timeliness coefficient of the information association problem point is determined according to the ratio of the data update duration of the retrieval source to the preset update duration.

[0051] It can be understood that the target retrieval source is the data source with the largest matching coefficient and all potential problem association points belong to information association problem points.

[0052] In another possible embodiment, the method for determining the target retrieval source is: Based on the information association situation between the retrieval source and the potential problem points, determine the potential problem points where the retrieval source has supplementary retrieval information, and use them as information association problem points; According to the data update cycle of the retrieval source, determine the timeliness coefficients of different information association problem points; Determine the matching coefficient of the target retrieval source based on the product of the proportion of the number of the potential problem points corresponding to different information association problem points and the timeliness coefficient, and determine whether the data source is the target retrieval source based on the matching coefficient.

[0053] In another possible embodiment, the method for determining the target retrieval source is as follows: Based on the information association situation between the retrieval source and the potential problem points, determine the potential problem points where the retrieval source has supplementary retrieval information, and use them as information association problem points. According to the data volume of the supplementary retrieval information of different information association problem points, determine the information association coefficients corresponding to different information association problem points; According to the data update period of the retrieval source, determine the timeliness coefficients of different information association problem points; Based on the average value of the product of the information association coefficients and the timeliness coefficients of different information association problem points, determine the matching coefficient of the target retrieval source, and determine whether the data source is the target retrieval source based on the matching coefficient.

[0054] Further, the information association coefficient is determined according to the ratio of the data volume of the supplementary retrieval information of the information association problem point to the amount of information to be verified and processed for the information association problem point.

[0055] Optionally, the method for determining the target retrieval source is as follows: Based on the information association situation between the retrieval source and the potential problem points, determine the potential problem points where the retrieval source has supplementary retrieval information, and use them as information association problem points. When there are potential problem points of the retrieval source that do not belong to the information association problem points, determine that the retrieval source does not belong to the target retrieval source; When there are no potential problem points of the retrieval source that do not belong to the information association problem points: Obtain the average update period of the retrieval information of the retrieval source. When the average update period does not meet the requirements, determine that the retrieval source does not belong to the target retrieval source; When the average update period meets the requirements: According to the data update period of the retrieval source and the supplementary retrieval information of the information association problem points, determine the timeliness coefficients of different information association problem points. When the average value of the timeliness coefficients of different information association problem points does not meet the requirements, determine that the retrieval source does not belong to the target retrieval source; When the average value of the timeliness coefficients of different information association problem points meets the requirements: Obtain the number of information association problem points whose timeliness coefficients do not meet the requirements. When the number of information association problem points whose timeliness coefficients do not meet the requirements is greater than the preset number of association problem points, determine that the retrieval source does not belong to the target retrieval source; When the number of information association problem points where the time - effect coefficient does not meet the requirements is not greater than the preset number of association problem points: According to the data volume of supplementary retrieval information for different information association problem points, determine the information association coefficients for different information association problem points. When the average value of the information association coefficients for different information association problem points does not meet the requirements, it is determined that the retrieval source does not belong to the target retrieval source; When the average value of the information association coefficients for different information association problem points meets the requirements: Based on the average value of the product of the information association coefficients for different information association problem points and the time - effect coefficient, determine the matching coefficient of the target retrieval source, and based on the matching coefficient, determine whether the data source is the target retrieval source.

[0056] Specifically, as Figure 5 shown, it is determined that adjustment processing of the retrieval plan is required, specifically including: Based on the deviation situation of supplementary retrieval information of different potential risk points in the target data source, determine the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information; Based on the deviation situation between the associated report content and the supplementary retrieval information, determine the number of potential risk points of the associated report content with deviations; According to the number of potential risk points of the associated report content with deviations, determine whether adjustment processing of the retrieval plan is required.

[0057] Furthermore, when the number of potential risk points of the associated report content with deviations is greater than the preset risk point number threshold, it is determined that adjustment processing of the retrieval plan is required.

[0058] It should be noted that when adjustment processing of the retrieval plan is not required, the potential risk points and the supplementary retrieval information of the potential risk points in the target data source are used as input quantities, and the evaluation conclusion of the credit report is determined using the output result of the large - model.

[0059] It can be understood that the evaluation conclusion of the credit report includes passing and not passing.

[0060] In another possible embodiment, it is determined that adjustment processing of the retrieval plan is required, specifically including: S41 Based on the deviation situation of supplementary retrieval information of different potential risk points in the target data source, determine the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information. Based on the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information, determine the data deviation coefficient of different potential risk points; S42 determines the risk correlation coefficient of the target data source at the potential risk point based on the data volume and data update cycle of the supplementary retrieval information of the target data source at the potential risk point; S43 determines the data deviation probability evaluation value of the credit report based on the risk correlation coefficients and data deviation coefficients of different potential risk points, and determines whether adjustment processing of the retrieval plan is required according to the data deviation probability evaluation value.

[0061] Further, when the data deviation probability evaluation value is greater than the preset deviation probability threshold, it is determined that no adjustment processing of the retrieval plan is required.

[0062] Optionally, the above step S41 includes the following contents: S411 determines the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information based on the deviation situation of the supplementary retrieval information of different potential risk points in the target data source. When there is no associated report content that deviates from the supplementary retrieval information, it is determined that no adjustment processing of the retrieval plan is required. When there is associated report content that deviates from the supplementary retrieval information, it proceeds to step S412; S412 obtains the quantity of the associated report content that deviates from the supplementary retrieval information. When the quantity of the associated report content that deviates from the supplementary retrieval information does not meet the requirements, it is determined that adjustment processing of the retrieval plan is required. When the quantity of the associated report content that deviates from the supplementary retrieval information meets the requirements, it proceeds to step S413; S413 determines the data deviation coefficients of different potential risk points based on the deviation situation between the associated report content of different potential risk points in the credit report and the supplementary retrieval information. When the sum of the data deviation coefficients of different potential risk points does not meet the requirements, it is determined that adjustment processing of the retrieval plan is required. When the sum of the data deviation coefficients of different potential risk points meets the requirements, it proceeds to step S414; S414 When there are potential risk points with data deviation coefficients that do not meet the requirements, it proceeds to step S415. When there are no potential risk points with data deviation coefficients that do not meet the requirements, it proceeds to step S42; S415 When the quantity of potential risk points with data deviation coefficients that do not meet the requirements is greater than the preset value of the quantity of risk points, it is determined that adjustment processing of the retrieval plan is required. When the quantity of potential risk points with data deviation coefficients that do not meet the requirements is not greater than the preset value of the quantity of risk points, it proceeds to step S42.

[0063] Optionally, the above step S42 includes the following contents: S421 determines the risk correlation coefficient of the target data source at the potential risk point based on the data volume and data update cycle of the supplementary retrieval information of the target data source at the potential risk point, and takes the potential risk points with a risk correlation coefficient less than the preset risk correlation coefficient threshold as shallow correlation risk points. When the number of shallow correlation risk points or the proportion of shallow correlation risk points in the number of potential risk points does not meet the requirements, it is determined that adjustment processing of the retrieval scheme is required. When both the number of shallow correlation risk points and the proportion of shallow correlation risk points in the number of potential risk points meet the requirements, step S422 is entered; S422 When the number of shallow correlation risk points is within the preset range of the number of shallow correlation risk points, step S423 is entered. When the number of shallow correlation risk points is not within the preset range of the number of shallow correlation risk points, step S424 is entered; S423 When the average value of the data deviation coefficients of the shallow correlation risk points is within the preset deviation coefficient range, it is determined that adjustment processing of the retrieval scheme is required. When the average value of the data deviation coefficients of the shallow correlation risk points is not within the preset deviation coefficient range, step S424 is entered; S424 determines the data source deviation amounts of different potential risk points based on the product of the risk correlation coefficient and the data deviation coefficient of the target data source at the potential risk point. When the average value of the data source deviation amounts of different potential risk points does not meet the requirements, it is determined that adjustment processing of the retrieval scheme is required. When the average value of the data source deviation amounts of different potential risk points meets the requirements, step S43 is entered.

[0064] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the description of the method embodiments.

[0065] The above describes specific embodiments of this specification. In some cases, the recorded actions or steps can be executed in a different order from that in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0066] The above description is only for one or more embodiments of this specification and is not intended to limit this specification. For those skilled in the art, various changes and modifications can be made to one or more embodiments of this specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification shall be included within the scope of the claims of this specification.

Claims

1. A micro-credit expert system based on a large language model, characterized in that: Specifically include: Problem point recording module, potential problem point output module, retrieval processing module, conclusion output module; The problem point recording module is responsible for recording the problem points of the credit report during the approval process; The potential problem point output module is responsible for identifying and processing the credit report and the problem points of the credit report using a large language model to obtain potential problem points; The retrieval processing module performs retrieval processing based on the potential problem points to obtain supplementary retrieval information of the potential problem points; The conclusion output module is responsible for outputting the evaluation conclusion of the credit report based on potential problem points and corresponding supplementary search information.

2. The micro-credit expert system based on a large language model as claimed in claim 1, characterized in that: The problem points include credit data, credit usage data, production and operation indicators, and financial indicators.

3. The micro-credit expert system based on a large language model as claimed in claim 1, characterized in that: The supplementary search information is a search result of data related to the potential problem point.

4. The micro-credit expert system based on a large language model as claimed in claim 1, characterized in that: Output of credit report assessment conclusions, including: The potential risk points and the supplementary search information of the potential risk points are used as input, and the output results of the large model are used to determine the evaluation conclusion of the credit report.

5. A small and micro loan approval method based on a large language model, applied to a small and micro loan expert system based on a large language model as described in any one of claims 1 to 4, characterized in that: Specifically include: Record the problem points of the credit report during the approval process, determine the related report content of different problem points according to the type of the problem points, and proceed to the next step when it is determined that the risk of the credit report meets the requirements based on the related report content of different problem points; Based on the credit report and the problem points of the credit report, a large language model is used to perform identification processing to obtain potential problem points, and a target search source is determined according to the information association and update status of different search sources and potential problem points; Determine the deviation of the supplementary search information of different potential risk points based on the search processing results of the target search source, and when it is determined that the search plan needs to be adjusted based on the deviation, obtain the supplementary search information of the credit report based on the full amount of search sources; An evaluation conclusion of the credit report is output based on the potential problem points and the corresponding supplementary search information.

6. The micro-credit approval method based on a large language model as claimed in claim 5, characterized in that: The problem points in the approval process are determined based on the problems in the approval process of the relevant personnel of the credit report.

7. The micro-credit approval method based on a large language model as claimed in claim 5, characterized in that: Determine that the risk of the credit report meets the requirements, including: Determining the associated report contents in different subunits of the credit report based on the associated report contents of the problem points; Determining the problem subunit of the associated report content according to the constituent data of the associated report content in different subunits; Whether the risk of the credit report meets the requirement is determined by the number of the problem sub-units.

8. The micro-credit approval method based on a large language model as claimed in claim 5, characterized in that: When the risk of the credit report does not meet the requirements, supplementary search information of the credit report is obtained based on all indicators and all search sources.

9. The micro-credit approval method based on a large language model as claimed in claim 5, characterized in that: Determine the need to adjust the search plan, including: Determine the deviation between the supplementary search information and the associated report content of the different potential risk points in the credit report based on the deviation of the supplementary search information of the target data source; Based on the deviation between the associated report content and the supplementary search information, determine the number of potential risk points of the associated report content with deviations; Determine whether the retrieval scheme needs to be adjusted based on the number of potential risk points in the associated report content with deviations.

10. The micro-credit approval method based on a large language model as claimed in claim 9, characterized in that: When there is no need to adjust the retrieval scheme, the potential risk points and the supplementary retrieval information of the potential risk points in the target data source are used as input, and the output results of the large model are used to determine the evaluation conclusion of the credit report.

Images