Visual file generation method and related device
By extracting the text content of the cooperation agreement between insurance companies and intermediary companies and using large language models to generate data representations of the protocol model, the problems of low manual processing efficiency and high probability of error in traditional solutions are solved, and efficient and accurate protocol information extraction and Visual file generation.
Patent Information
- Application Number
- CN202411505658.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2025-02-18
AI Technical Summary
In traditional solutions, the management of cooperation agreements between insurance companies and intermediary companies relies on manual reading and entry, resulting in low efficiency and high probability of errors.
By determining the target protocol, extracting its text content, and generating a data representation of the protocol model using a pre-trained large language model, a visual file is generated based on the data representation.
It realizes the extraction of key information in the cooperation agreement with high efficiency and accuracy, and generates visual files, solving the problems of low manual processing efficiency and high probability of errors in traditional solutions.
Smart Images
Figure CN119474360A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of protocol management, and in particular to a method for generating a visualization file and a related device. Background Art
[0002] In the process of cooperation between insurance companies and intermediary companies, a large number of cooperation agreements will be generated. In the traditional solution, the management process of cooperation agreements is that the agreement manager manually reads the agreement and enters the key information in the agreement into the system one by one to achieve the management of the cooperation agreement. Since the traditional solution relies heavily on manual processing of cooperation agreements, there are problems of low efficiency and high probability of errors. Therefore, how to efficiently and accurately extract the key information in the cooperation agreement and generate a visual file has become one of the technical problems that need to be solved in the field of agreement management technology. Summary of the invention
[0003] Based on the above problems, the present application provides a method for generating a visualization file, which is used to extract key information in a cooperation agreement efficiently and accurately and generate a visualization file.
[0004] The embodiments of the present application disclose the following technical solutions:
[0005] The first aspect of the present application discloses a method for generating a visualization file, comprising:
[0006] Determine the target agreement; the target agreement is a cooperation agreement between the insurance institution and the intermediary institution;
[0007] Extracting text content in the target protocol;
[0008] Generate a data representation of a protocol model corresponding to the target protocol through a target macromodel according to the text content and the target prompt word template; the target macromodel is a large language model trained in advance and used to extract key information from the text content and generate a data representation of the protocol model; the key information includes cooperation period, cooperation mode, fee calculation method and payment rules;
[0009] A target visualization file is generated based on the data representation of the protocol model.
[0010] In an optional implementation, the target prompt word template includes: a prompt word, a JSON template structure and instructions for use; the JSON template structure is used to specify the data representation form of the protocol model; the instructions for use include a description of the protocol cycle format in the cooperation agreement, a description of the fee list format or a description of the renewal fee rate adjustment rules.
[0011] In an optional implementation, the step of training the target large model includes:
[0012] Obtain a sample agreement; the sample agreement is an agreement in the cooperation agreement between the insurance institution and the intermediary institution;
[0013] Extracting text content in the sample agreement as sample text content;
[0014] Inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model;
[0015] generating a prediction visualization file based on the data representation of the prediction protocol model;
[0016] Based on the sample protocol, the prediction visualization file is modified to obtain a sample visualization file;
[0017] Based on the difference between the predicted visualization file and the sample visualization file, the initial large model is trained to obtain the target large model.
[0018] In an optional implementation, the generating a target visualization file based on the data representation of the protocol model includes:
[0019] Storing the data representation of the protocol model in JSON format to obtain target JSON data;
[0020] The target JSON data is processed to generate the target visualization file.
[0021] In an optional implementation, if the format of the target protocol is a picture, extracting the text content in the target protocol is specifically:
[0022] The target protocol is processed by optical character recognition technology to obtain text content in the target protocol.
[0023] A second aspect of the present application discloses a device for generating a visualization file, the device comprising:
[0024] A file acquisition module, used to determine a target agreement; the target agreement is a cooperation agreement between an insurance institution and an intermediary institution;
[0025] A text recognition module, used to extract text content in the target protocol;
[0026] A data representation generation module is used to generate a data representation of a protocol model corresponding to the target protocol according to the text content and the target prompt word template through a target large model; the target large model is a large language model trained in advance and used to extract key information from the text content and generate a data representation of the protocol model; the key information includes a cooperation period, a cooperation method, a handling fee calculation method and a payment rule;
[0027] The visualization file generation module is used to generate a target visualization file based on the data representation of the protocol model.
[0028] In an optional implementation, the device further includes a target large model training module, and the target large model training module specifically includes:
[0029] A sample agreement acquisition unit, used to acquire a sample agreement; the sample agreement is an agreement in the cooperation agreement between the insurance institution and the intermediary institution;
[0030] A sample text content extraction unit, used to extract text content in the sample agreement as sample text content;
[0031] A prediction data representation generating unit, used for inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model;
[0032] A prediction visualization file generating unit, configured to generate a prediction visualization file based on the data representation of the prediction protocol model;
[0033] A sample visualization file generating unit, configured to modify the prediction visualization file based on the sample protocol to obtain a sample visualization file;
[0034] The target large model acquisition unit is used to train the initial large model based on the difference between the predicted visualization file and the sample visualization file to obtain the target large model.
[0035] In an optional implementation, the visualization file generation module includes:
[0036] A JSON file acquisition unit, used to store the data representation of the protocol model in JSON format to obtain target JSON data;
[0037] The target visualization file generating unit is used to process the target JSON data to generate the target visualization file.
[0038] A third aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method introduced in any implementation manner of the first aspect.
[0039] A fourth aspect of the present application provides an electronic device, including:
[0040] a memory having a computer program stored thereon;
[0041] A processor is used to execute the computer program in the memory to implement the steps of the method introduced in any implementation manner of the first aspect.
[0042] Compared with the prior art, this application has the following beneficial effects:
[0043] The present application discloses a method for generating a visualization file, including: determining a target agreement from a cooperation agreement between an insurance institution and an intermediary institution; extracting text content from the target agreement; inputting the above text content and a target prompt word template into a target macro model to generate a data representation of a protocol model corresponding to the target agreement; and generating a target visualization file based on the data representation of the protocol model. In this way, after obtaining the text content in the target agreement, a data representation of a protocol model containing key information such as a cooperation period, cooperation method, fee billing method, and payment rules can be directly generated through the target macro model, thereby solving the problem of low efficiency and high error probability in the traditional solution of manually extracting key information from the agreement; based on the data representation of the generated protocol template, a target visualization file can be generated. Therefore, the solution in the present application can extract key information from the cooperation agreement with high efficiency and accuracy, and generate a visualization file. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0045] Figure 1 A flowchart of a method for generating a visualization file provided in an embodiment of the present application;
[0046] Figure 2 A schematic diagram of a JSON template structure provided for an embodiment of the present application;
[0047] Figure 3 A schematic diagram of an instruction for use provided in an embodiment of the present application;
[0048] Figure 4 A schematic diagram of data representation of a protocol model provided in an embodiment of the present application;
[0049] Figure 5 A flowchart of a method for obtaining a target large model provided in an embodiment of the present application;
[0050] Figure 6 A schematic diagram of a prediction visualization file provided in an embodiment of the present application;
[0051] Figure 7 A schematic diagram of adjusting parameters of a prediction visualization file provided in an embodiment of the present application;
[0052] Figure 8 A schematic diagram of the structure of a device for generating a visualization file provided in an embodiment of the present application;
[0053] Fig. 9 A schematic diagram of the structure of another device for generating a visualization file provided in an embodiment of the present application. DETAILED DESCRIPTION
[0054] In the process of cooperation between insurance companies and intermediary companies, a large number of cooperation agreements will be generated. Currently, there are two traditional solutions for managing cooperation agreements. The first traditional solution is that the agreement manager manually reads the agreement and enters the key information in the agreement into the system one by one. The second traditional solution is that the agreement manager organizes the agreement into an Excel spreadsheet using a pre-set template, and then imports the Excel spreadsheets in a unified format into the system in batches.
[0055] Although the second traditional solution reduces the workload of manual data entry to a certain extent, the agreement manager still needs to engage in complex information conversion and sorting work, and there are still problems of low work efficiency and high error probability. Therefore, how to extract key information in the cooperation agreement efficiently and accurately and generate visual files has become one of the technical problems that need to be solved in the field of agreement management technology.
[0056] The present application discloses a method for generating a visualization file, including: determining a target agreement from a cooperation agreement between an insurance institution and an intermediary institution; extracting text content from the target agreement; inputting the above text content and a target prompt word template into a target macro model to generate a data representation of a protocol model corresponding to the target agreement; and generating a target visualization file based on the data representation of the protocol model. In this way, after obtaining the text content in the target agreement, a data representation of a protocol model containing key information such as a cooperation period, cooperation method, fee billing method, and payment rules can be directly generated through the target macro model, thereby solving the problem of low efficiency and high error probability in the traditional solution of manually extracting key information from the agreement; based on the data representation of the generated protocol template, a target visualization file can be generated. Therefore, the solution in the present application can extract key information from the cooperation agreement with high efficiency and accuracy, and generate a visualization file.
[0057] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0058] Figure 1 A flowchart of a method for generating a visualization file provided in an embodiment of the present application. Figure 1 As shown, the method for generating a visualization file disclosed in this application includes:
[0059] S101, determine the target protocol.
[0060] The target agreement in this application is any one of a large number of cooperation agreements between insurance institutions and intermediary institutions to be processed.
[0061] It is understandable that the cooperation agreement between insurance companies and intermediary institutions contains a large amount of relevant content regarding the cooperation period, cooperation mode, fee calculation method and payment rules; these contents are key information in the cooperation agreement.
[0062] S102: extract text content in the target protocol.
[0063] The cooperation agreement between the insurance company and the intermediary company is in the form of a PDF file or an image file. If the target agreement is a PDF file, the text content in the PDF file can be directly extracted; if the target file is an image file, the text content in the target agreement can be extracted through optical character recognition technology.
[0064] It should be noted that, since extracting text content from a PDF file is well known to those skilled in the art, this content will not be further elaborated in this application.
[0065] S103, generating a data representation of a protocol model corresponding to the target protocol according to the text content and the target prompt word template through the target macro model.
[0066] The target large model in this application is a large language model that is pre-trained and used to extract key information such as cooperation period, cooperation mode, fee calculation method and payment rules from text content and generate data representation of the agreement model. The training process of the target large language model will be described in detail in the subsequent embodiments of this application.
[0067] The target prompt word template in this application includes: prompt words, JSON template structure and instructions for use.
[0068] Among them, the prompt words prompt the target large language model to pay attention to the information and problems that need to be solved.
[0069] Exemplarily, the prompt words in the target prompt word template may be: This agreement model is used to describe key terms in insurance brokerage business cooperation, especially regarding the cooperation period, handling fee rate (commission rate) and the impact of the continuation rate on the handling fee rate.
[0070] Among them, the JSON template structure is used to specify the data representation format of the protocol model.
[0071] Figure 2 A schematic diagram of a JSON template structure provided in an embodiment of the present application. Figure 2 As shown, the template structure specifies the start and end dates of the agreement period; information such as the renewal period, age group, first-year fee rate and subsequent annual fee rate of the fee rate list; and information such as the 12-month premium continuation rate, threshold and adjustment value of the renewal fee rate adjustment rule, and the 25-month premium continuation rate, threshold, adjustment value and adjustment coefficient.
[0072] The instructions for use include an explanation of the format of the agreement cycle in the cooperation agreement, an explanation of the format of the handling fee list, or an explanation of the renewal handling fee rate adjustment rules.
[0073] Figure 3 A schematic diagram of a method for using the embodiment of the present application is provided. Figure 3 As shown, the usage rate provides detailed explanations of the formats of the agreement start date and agreement end date in the agreement period; the payment period and usage age group in the fee rate list; and the rules for adjusting the renewal fee rate.
[0074] The text content and the target prompt word template in S102 are simultaneously input into the target large language model. The target large language model extracts key information from the text content, including but not limited to cooperation period, cooperation mode, fee calculation method and payment rules, and generates a data representation of the protocol model corresponding to the target protocol.
[0075] Figure 4 A schematic diagram of data representation of a protocol model provided in an embodiment of the present application. Figure 4 As shown, the data representation of the agreement model gives the start date of the agreement period as November 8, 2023, and the end date as November 7, 2024; the annual fee period in the handling fee list is 3 years, the age group is 45 years old and below, and the first year handling fee rate is 0.29, etc., which are the key information that needs to be paid attention to in the target agreement.
[0076] S104: Generate a target visualization file based on the data representation of the protocol model.
[0077] After obtaining the data representation of the protocol model corresponding to the target protocol, the data representation of the protocol model is stored in JSON format to obtain JSON data; then the target JSON data is processed to generate the target visualization file.
[0078] It should be noted that since parsing JSON data and generating a target visualization file is well known to those skilled in the art, this content will not be further elaborated in this application.
[0079] The target visualization file in this application is a table. Those skilled in the art can also display the target visualization file to the user in the form of a legend. The specific form of the target visualization file is not limited in this application.
[0080] After obtaining the target visualization file, the target visualization file is stored in the target location to achieve the management of the target protocol.
[0081] In summary, the present application discloses a method for generating a visualization file, including: determining a target agreement from a cooperation agreement between an insurance institution and an intermediary institution; extracting text content from the target agreement; inputting the above text content and the target prompt word template into a target macro model to generate a data representation of a protocol model corresponding to the target agreement; based on the data representation of the protocol model, generating a target visualization file. In this way, after obtaining the text content in the target agreement, the data representation of the protocol model containing key information such as the cooperation cycle, cooperation method, fee billing method, and payment rules can be directly generated through the target macro model, solving the problem of low efficiency and high error probability in the traditional solution of manually extracting key information from the agreement; based on the data representation of the generated protocol template, a target visualization file can be generated. Therefore, the solution in the present application can extract key information from the cooperation agreement with high efficiency and accuracy, and generate a visualization file.
[0082] Figure 5 A flowchart of a method for obtaining a target large model provided in an embodiment of the present application. Figure 5 As shown, the method for obtaining the target large model in this application, that is, the training steps of the target large model include:
[0083] S501, obtaining a sample protocol.
[0084] Among them, the sample agreement is the agreement in the cooperation agreement between the insurance institution and the intermediary institution.
[0085] S502: extract text content in the sample agreement as sample text content.
[0086] S503, inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model.
[0087] S504: Generate a prediction visualization file based on the data representation of the prediction protocol model.
[0088] Among them, the contents of S502-S504 refer to the relevant introduction of S102-S104, which will not be repeated here.
[0089] It should be emphasized that the specific form of the initial large model is not limited in the present application, and those skilled in the art may adopt Wenxinyiyan or Text-to-Text Transfer Transformer as the initial large model.
[0090] S505: Based on the sample protocol, modify the prediction visualization file to obtain a sample visualization file.
[0091] It is understandable that based on the data representation of the prediction protocol model generated by the initial large model, there may be problems or errors in the generated prediction visualization file.
[0092] Figure 6 A schematic diagram of a prediction visualization file provided in an embodiment of the present application. Figure 6 As shown, there may be a situation where the feature value under the feature parameter in the prediction visualization file is inconsistent with the feature value under the feature parameter in the sample protocol.
[0093] For example, in the forecast visualization file, the insurance with the insurance type code A00001 in the basic handling fee rate has a first-year rate of 60%, which is different from the first-year rate of 50% for the insurance A00001 recorded in the sample agreement.
[0094] In order to solve problems or errors in the prediction visualization file, it is necessary to refer to the sample protocol and verify the characteristic values under each characteristic parameter in the prediction visualization file until the characteristic values under each characteristic parameter in the corrected prediction visualization file are ensured to be consistent with the characteristic values under each characteristic parameter in the sample protocol, so as to obtain the sample visualization file.
[0095] Figure 7 A schematic diagram of adjusting parameters of a prediction visualization file provided in an embodiment of the present application. Figure 7 As shown, after adjusting the characteristic values under the characteristic parameters in the prediction visualization file, a sample visualization file can be obtained; the manually corrected content is displayed in red.
[0096] S506: Based on the difference between the predicted visualization file and the sample visualization file, the initial large model is trained to obtain the target large model.
[0097] After obtaining the sample visualization file, the parameters of the initial large model are adjusted based on the difference between the sample visualization file and the predicted visualization file until the difference between the sample visualization file and the predicted visualization file meets a preset difference threshold, thereby obtaining a trained target large model.
[0098] Based on the visualization file generation method disclosed in the aforementioned embodiment, the present application also discloses a visualization file generation device. Figure 8 A schematic diagram of a structure of a device for generating a visualization file provided in an embodiment of the present application. Figure 8 As shown, the visualization file generation device 800 disclosed in the present application includes:
[0099] The file acquisition module 801 is used to determine the target agreement; the target agreement is the cooperation agreement between the insurance institution and the intermediary institution;
[0100] A text recognition module 802, used to extract text content in the target protocol;
[0101] The data representation generation module 803 is used to generate the data representation of the protocol model corresponding to the target protocol according to the text content and the target prompt word template through the target large model; the target large model is a large language model trained in advance and used to extract key information from the text content and generate the data representation of the protocol model; the key information includes the cooperation period, cooperation mode, fee calculation method and payment rules;
[0102] The visualization file generating module 804 is used to generate a target visualization file based on the data representation of the protocol model.
[0103] In an optional implementation, the visualization file generation device 800 further includes a target large model training module, and the target large model training module specifically includes:
[0104] A sample agreement acquisition unit, used to acquire a sample agreement; the sample agreement is an agreement in the cooperation agreement between the insurance institution and the intermediary institution;
[0105] A sample text content extraction unit, used to extract text content in the sample agreement as sample text content;
[0106] A prediction data representation generating unit, used for inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model;
[0107] A prediction visualization file generating unit, configured to generate a prediction visualization file based on the data representation of the prediction protocol model;
[0108] A sample visualization file generating unit, configured to modify the prediction visualization file based on the sample protocol to obtain a sample visualization file;
[0109] The target large model acquisition unit is used to train the initial large model based on the difference between the predicted visualization file and the sample visualization file to obtain the target large model.
[0110] In an optional implementation, the visualization file generation module 804 includes:
[0111] A JSON file acquisition unit, used to store the data representation of the protocol model in JSON format to obtain target JSON data;
[0112] The target visualization file generating unit is used to process the target JSON data to generate the target visualization file.
[0113] In an optional implementation, the text recognition module 802 includes:
[0114] The optical character recognition unit is used to process the target protocol through optical character recognition technology to obtain the text content in the target protocol.
[0115] Based on the visualization file generation method disclosed in the aforementioned embodiment, the present application also discloses a visualization file generation device. Fig. 9 A schematic diagram of the structure of another device for generating a visualization file provided in an embodiment of the present application. Fig. 9 As shown, the visualization file generation device includes a file upload module 901, a text recognition module 902, a large model 903 (i.e., the initial large model in this application), a data visualization processing module 904, a protocol storage module 905 and a model training module 906.
[0116] The file upload module 901 is used to receive target protocols uploaded by users; it supports target protocols in image format and PDF format.
[0117] The text recognition module 902 is used to parse the uploaded protocol file (i.e., the uploaded target protocol). If the target protocol is in PDF format, the PDF parsing program is directly called to obtain the text content in the target protocol; if the target protocol is in image format, the optical character recognition technology is used to extract the text content in the image; finally, the text content in the target protocol is extracted.
[0118] The large model 903 is used to parse the text content and identify the key information in the protocol; through the target prompt word template, the key information is converted into the data representation of the predicted protocol model, and stored in the JSON format to obtain the predicted JSON data.
[0119] The data visualization processing module 904 is used to parse the prediction JSON data, generate a prediction visualization file, and display it visually on the front-end page; the user can intuitively see the key information of the protocol on the front-end page; the results of the system analysis can be verified and proofread on the visualization interface; the user can manually modify the parts with recognition errors; if the recognition is correct or all modifications are correct, the user can click the "Submit" button to store the corrected prediction visualization file in the protocol storage module.
[0120] The protocol storage module 905 is used to store the predicted visualization file after manual correction, that is, to store the target visualization file.
[0121] The model training module 906 is used to train and fine-tune the large model according to the difference between the corrected target visualization file and the predicted visualization file generated by the large model, so as to improve the parsing accuracy of the large model. After stopping the training, the target large model in this application is obtained.
[0122] pass Fig. 9 The visual file generation device in can realize the automatic parsing and input of the target agreement, which significantly improves the efficiency and accuracy of agreement management. The large model learns, proofreads, and fine-tunes at work, forming a positive cycle, with better and better results and higher and higher recognition accuracy. The solution in this application not only reduces the workload of managers, but also improves the efficiency of identifying key information in cooperation agreements, reduces the probability of human errors, and can provide efficient and reliable technical support for the cooperation between insurance companies and intermediary companies.
[0123] Based on the visualization file generation method and device provided in the aforementioned embodiments, the present application also provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements some or all of the steps in the visualization file generation method mentioned above.
[0124] Based on the method and device for generating a visualization file provided in the foregoing embodiments, the present application further provides an electronic device, comprising: a memory having a computer program stored thereon;
[0125] The processor is used to execute the computer program in the memory to implement part or all of the steps in the method for generating a visualization file provided in the above embodiment.
[0126] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described above is merely schematic, in which the unit described as a separate component may or may not be physically separated, and the component prompted as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.
[0127] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for generating a visualization file, characterized in that: The method comprises: Determine the target agreement; the target agreement is a cooperation agreement between the insurance institution and the intermediary institution; Extracting text content in the target protocol; Generate a data representation of a protocol model corresponding to the target protocol through a target macromodel according to the text content and the target prompt word template; the target macromodel is a large language model trained in advance and used to extract key information from the text content and generate a data representation of the protocol model; the key information includes cooperation period, cooperation mode, fee calculation method and payment rules; A target visualization file is generated based on the data representation of the protocol model.
2. The method according to claim 1, characterized in that The target prompt word template includes: prompt words, JSON template structure and instructions for use; the JSON template structure is used to specify the data representation form of the protocol model; the instructions for use include a description of the protocol cycle format in the cooperation agreement, a description of the fee list format or a description of the renewal fee rate adjustment rules.
3. The method according to claim 1 or 2, characterized in that: The training steps of the target large model include: Obtain a sample agreement; the sample agreement is an agreement in the cooperation agreement between the insurance institution and the intermediary institution; Extracting text content in the sample agreement as sample text content; Inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model; generating a prediction visualization file based on the data representation of the prediction protocol model; Based on the sample protocol, the prediction visualization file is modified to obtain a sample visualization file; Based on the difference between the predicted visualization file and the sample visualization file, the initial large model is trained to obtain the target large model.
4. The method according to claim 1 or 2, characterized in that: The data representation based on the protocol model generates a target visualization file, including: Storing the data representation of the protocol model in JSON format to obtain target JSON data; The target JSON data is processed to generate the target visualization file.
5. The method according to claim 1 or 2, characterized in that: If the format of the target protocol is a picture, extracting the text content in the target protocol is specifically: The target protocol is processed by optical character recognition technology to obtain text content in the target protocol.
6. A device for generating a visualization file, characterized in that: The device comprises: A file acquisition module, used to determine a target agreement; the target agreement is a cooperation agreement between an insurance institution and an intermediary institution; A text recognition module, used to extract text content in the target protocol; A data representation generation module is used to generate a data representation of a protocol model corresponding to the target protocol according to the text content and the target prompt word template through a target large model; the target large model is a large language model trained in advance and used to extract key information from the text content and generate a data representation of the protocol model; the key information includes a cooperation period, a cooperation method, a handling fee calculation method and a payment rule; The visualization file generation module is used to generate a target visualization file based on the data representation of the protocol model.
7. The device according to claim 6, characterized in that The device also includes a target large model training module, and the target large model training module specifically includes: A sample agreement acquisition unit, used to acquire a sample agreement; the sample agreement is an agreement in the cooperation agreement between the insurance institution and the intermediary institution; A sample text content extraction unit, used to extract text content in the sample agreement as sample text content; A prediction data representation generating unit, used for inputting the sample text content and the target prompt word template into the initial large model to generate a data representation of the prediction protocol model; A prediction visualization file generating unit, configured to generate a prediction visualization file based on the data representation of the prediction protocol model; A sample visualization file generating unit, configured to modify the prediction visualization file based on the sample protocol to obtain a sample visualization file; The target large model acquisition unit is used to train the initial large model based on the difference between the predicted visualization file and the sample visualization file to obtain the target large model.
8. The method according to claim 6, characterized in that The visualization file generation module includes: A JSON file acquisition unit, used to store the data representation of the protocol model in JSON format to obtain target JSON data; The target visualization file generating unit is used to process the target JSON data to generate the target visualization file.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 5 are implemented.
10. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 5.