Personalized and interactive credit report generation method and device

Abstract

The invention relates to a personalized and interactive credit report generation method and device, and the method comprises the following steps: obtaining evaluation data of an evaluation object, the evaluation data comprising financial information and credit related information; generating a preliminary credit evaluation report based on the evaluation data by using a natural language processing technology; receiving feedback data, judging whether the feedback data is report confirmation information or not, if not, adjusting the preliminary credit evaluation report in a personalized manner based on the feedback data, generating an updated credit evaluation report, and repeatedly receiving the feedback data until the feedback data is the report confirmation information, and outputting the current credit evaluation report as a final credit evaluation report. Compared with the prior art, the method has the advantages of improving report generation efficiency and user experience and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and device for generating a personalized and interactive credit report. Background Art

[0002] Credit reports (also commonly referred to as credit reports or credit investigation reports) play a vital role in modern society and are the main basis for financial institutions to assess the credit risk of individuals or enterprises. Through credit reports, financial institutions can more accurately assess credit risks and thus formulate reasonable credit policies.

[0003] The credit report writing process generally includes collecting financial data, analyzing financial status, assessing credit risk, and writing structured reports. Each step is time-consuming, and the writers often face significant time pressure. Traditionally, these steps rely heavily on manual operations, which increases time costs and the risk of errors. Existing technologies are often inefficient when processing large amounts of complex data, and are also deficient in user interaction and personalized services, requiring a lot of manpower investment.

[0004] In order to improve the efficiency of credit report writing, existing technologies have introduced more traditional statistical models and credit scoring algorithms, but most of these methods are based on fixed financial indicators and historical data, do not consider personalized services, and have poor user experience. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a personalized and interactive credit report generation method and device that improves report generation efficiency and user experience.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A personalized and interactive credit report generation method, comprising the following steps:

[0008] Obtaining assessment data of the assessment object, wherein the assessment data includes financial information and credit related information;

[0009] Using natural language processing technology, generate a preliminary credit assessment report based on the assessment data;

[0010] Receive feedback data, determine whether the feedback data is report confirmation information, and if not, adjust the preliminary credit assessment report based on the feedback data, generate an updated credit assessment report, repeatedly receive feedback data until the feedback data is report confirmation information, and output the current credit assessment report as the final credit assessment report.

[0011] Furthermore, the financial information includes financial statements, historical loan records, deposit information and financial product holdings.

[0012] Furthermore, the credit-related information includes credit rating and market reputation.

[0013] Furthermore, a large language model is used to generate the preliminary credit assessment report.

[0014] Furthermore, the large language model is pre-trained using large-scale financial field text data as training data, and the credit report data is used to fine-tune the model based on the pre-trained model.

[0015] Furthermore, based on a set update cycle, the large language model is retrained regularly.

[0016] Furthermore, the feedback data includes error correction information, additional provided information and / or analysis result objection information.

[0017] Furthermore, the personalized adjustment specifically includes:

[0018] Utilize natural language processing technology to analyze the feedback data, understand the results based on the context, and determine whether there are data errors in the current credit assessment report. If so, verify the data and correct the errors;

[0019] According to the analysis result of the feedback data, determining whether the report content needs to be adjusted, and if so, performing report adjustment processing based on the analysis result of the feedback data;

[0020] Determine whether the adjusted report passes the quality review. If so, generate an updated credit assessment report.

[0021] Furthermore, the quality review includes policy and risk review and risk reassessment.

[0022] The present invention also provides a personalized and interactive credit report generating device, comprising the following steps:

[0023] A data acquisition module, used to acquire the evaluation data of the evaluation object, wherein the evaluation data includes financial information and credit-related information;

[0024] A preliminary assessment module, used to generate a preliminary credit assessment report based on the assessment data using natural language processing technology;

[0025] The personalization engine module is connected to the interactive user interface and is used to receive feedback data transmitted by the interactive user interface, determine whether the feedback data is report confirmation information, and if not, adjust the preliminary credit assessment report based on the feedback data, generate an updated credit assessment report, repeatedly receive feedback data until the feedback data is report confirmation information, and output the current credit assessment report as the final credit assessment report.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. The present invention can provide more in-depth and personalized financial analysis by integrating advanced natural language processing technology and a personalized adjustment engine. At the same time, it improves the user experience and the convenience of data submission through interactive user feedback. The automatically generated credit report greatly reduces the heavy workload of bank account managers, liberates manpower, and greatly improves the efficiency of credit report writing.

[0028] 2. By automatically collecting the company's financial data and bank history records, it reduces manual input errors and time expenditures. At the same time, it can process multiple data formats and batch data, improving the efficiency and accuracy of data collection.

[0029] 3. The present invention is provided with a personalized adjustment engine, which can adjust reports according to feedback data, including being able to adjust assessment rules according to different loan products to provide more customized assessment results. The feedback mechanism ensures continuous optimization of the assessment strategy and self-adjustment of the model.

[0030] 4. The system of the present invention is provided with an intuitive and responsive interactive user interface, which allows enterprises to easily submit data and view evaluation results in real time, and can enhance user interaction experience and satisfaction through real-time feedback and notifications.

[0031] 5. The present invention provides a safe, efficient and customizable enterprise credit assessment solution by combining advanced technology and customer-friendly design, which not only improves the operational efficiency of enterprises, but also improves the satisfaction of enterprise customers, and promotes risk management and market competitiveness of banking business. Account managers can generate credit reports in a shorter time and have more time for high-value tasks such as customer relationship management and strategic planning. In addition, the application of automation and intelligent tools can also help reduce the error rate and improve the quality of reports, thereby potentially reducing credit risks caused by reporting errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the process of the present invention;

[0033] Figure 2 This is a flow chart of large language model training and deployment in the present invention;

[0034] Figure 3 The flowchart of the personalized adjustment engine of the present invention;

[0035] Figure 4 It is a system schematic diagram of the present invention. DETAILED DESCRIPTION

[0036] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0037] Example 1

[0038] The present embodiment provides a personalized and interactive credit report generation method, comprising the following steps: obtaining assessment data of an assessment object, wherein the assessment data includes financial information and credit-related information; using natural language processing technology to generate a preliminary credit assessment report based on the assessment data; receiving feedback data, and determining whether the feedback data is report confirmation information; if not, adjusting the preliminary credit assessment report based on the feedback data in a personalized manner, generating an updated credit assessment report, repeatedly receiving feedback data until the feedback data is report confirmation information, and outputting the current credit assessment report as the final credit assessment report.

[0039] refer to Figure 1 As shown, the process of a bank account manager using the above-mentioned personalized and interactive credit report generation method to generate a credit assessment report may include the following steps:

[0040] 1. Data collection stage:

[0041] Automatically obtain the assessment data of the assessment object, which includes financial information and credit-related information.

[0042] Financial information includes the company's financial statements, such as balance sheets, income statements, cash flow statements, etc., as well as the company's historical bank loan records, deposits and other financial products. In addition to direct financial data, credit-related information is also collected, including corporate credit ratings, market reputation and other credit indicators.

[0043] Specifically, the above data is obtained from the financial data and bank history records submitted by corporate customers on the client side.

[0044] 2. Data processing and draft report generation stage:

[0045] The collected data is analyzed using natural language processing technology to generate a preliminary credit assessment report.

[0046] In this embodiment, the preliminary credit assessment report is generated using a large language model, which is pre-trained using large-scale financial field text data as training data, and fine-tuned using the credit report data based on the pre-trained model.

[0047] Specifically, this embodiment uses the advanced natural language processing algorithm Transformer. The model is trained to understand and analyze financial and credit-related data. The model generates a structured credit report based on the understood data, which includes charts and text explanations, provides an in-depth analysis of the company's financial status, and liberates the manual labor of account managers.

[0048] In a preferred embodiment, the large language model is retrained periodically based on a set update cycle.

[0049] 3. Report personalization adjustment stage:

[0050] Receive feedback data, adjust the assessment model according to the bank's credit products and policies, and the specific needs of the enterprise, and generate an adjusted credit report. In this embodiment, the feedback data may include error correction information, additional information, analysis result objection information, etc.

[0051] Personalization has the following features:

[0052] Dynamic adjustment: Based on different loan types and corporate needs, banks can set specific assessment templates and risk parameters, and the engine will dynamically adjust the assessment logic accordingly.

[0053] Feedback optimization: Evaluation results are used as feedback to optimize future evaluation strategies and accuracy through continuous learning.

[0054] 4. Interactive evaluation and confirmation stage:

[0055] The adjusted report is presented to the corporate client, who can directly request further adjustments to the report or confirm the report on the platform.

[0056] 5. Customer feedback integration stage:

[0057] If the corporate customer requests further adjustments, make modifications based on the feedback and resubmit to the customer.

[0058] Once a corporate client confirms a report, it will be locked to prevent further modification.

[0059] 6. Credit report submission and loan processing stage:

[0060] Confirmed reports are automatically sent to account managers.

[0061] The account manager reviews the report and proceeds with the loan request based on the content of the report.

[0062] refer to Figure 2 As shown in the figure, the training and application process of the intelligent language processing engine, that is, the above-mentioned large language model, is as follows:

[0063] 1. Training process:

[0064] 1) Data preparation

[0065] Collection: Aggregate a large amount of text data in the financial field, including public reports, news, professional articles, historical credit reports, etc.

[0066] Cleaning: Clean the collected data, remove irrelevant information, and retain useful data.

[0067] Annotation: Annotate data to establish the relationship between the data and specific credit assessment tasks.

[0068] 2) Model selection

[0069] Architecture: Select a suitable language model architecture. This invention uses Transformer.

[0070] Configuration: Determine the model's size, number of layers, number of heads, and other parameter configurations.

[0071] 3) Pre-training

[0072] Language understanding: Train models on large-scale text data to learn the basic structure and meaning of language.

[0073] Task-agnostic: This step is usually not task-specific, but rather allows the model to gain broad language capabilities.

[0074] 4) Fine-tuning

[0075] Specific dataset: Fine-tune the model using a dataset related to credit reports.

[0076] Task Specialization: Train the model to understand financial terms, concepts, and related text structures.

[0077] 5) Verification and testing

[0078] Performance evaluation: Test the model performance on the validation set to ensure its accuracy and reliability.

[0079] Iterative optimization: Adjust model parameters based on test results and perform multiple rounds of iterations to improve performance.

[0080] 2. Application process:

[0081] 1) Deployment

[0082] Integration: Deploy the trained model to the bank’s cloud platform.

[0083] Interface: Develop an API or service interface so that other parts of the system can call the model.

[0084] 2) Real-time processing

[0085] Input: Receives text input from users or other parts of the system, such as financial report content, user queries, etc.

[0086] Prediction: Use the model to process input text and provide predicted output, such as generating reports, answering questions, etc.

[0087] 3) Interactive Adjustment

[0088] User feedback: Receive user feedback on model output.

[0089] Optimize output: Based on the feedback, adjust model parameters or output to improve personalization and accuracy.

[0090] 4) Continuous learning

[0091] Data accumulation: New data is accumulated during the operation of the system.

[0092] Retraining: Regularly retrain the model with new data to ensure that the model keeps pace with industry developments.

[0093] Through the above steps, the intelligent language processing engine (large language model) not only learns the basic rules of language, but also optimizes it specifically for the needs of the financial credit field. In actual applications, it can understand and generate professional credit assessment reports, and at the same time learn and adjust based on user feedback to provide more accurate and personalized services.

[0094] refer to Figure 3 As shown, the above-mentioned personalized adjustment engine is a key component designed to customize and optimize the credit assessment report. The specific working process of the personalized adjustment engine includes: using natural language processing technology to parse the feedback data, and judging whether there are data errors in the current credit assessment report based on the context understanding results, if so, correcting the erroneous content after data verification; judging whether the report content needs to be adjusted based on the parsing results of the feedback data, if so, adjusting the report based on the parsing results of the feedback data; judging whether the adjusted report passes the quality review, if so, generating an updated credit assessment report.

[0095] Specifically, the workflow is described in detail as follows:

[0096] 1) Receive preliminary credit assessment report and feedback

[0097] Input: The preliminary credit assessment report is generated by an intelligent language processing engine and is based on data provided by corporate customers (such as financial statements, historical loan records, etc.).

[0098] User Feedback: Enterprise customers provide feedback on the preliminary evaluation report through an interactive user interface, including error corrections, additional information provision, or objections to certain analysis results.

[0099] 2) Analyze and understand feedback

[0100] Parsing: The engine uses Natural Language Processing (NLP) technology to parse the text feedback provided by customers.

[0101] Contextual understanding: Determine which sections of the report the feedback is relevant to and understand the client’s specific needs.

[0102] 3) Data verification and update

[0103] Verification: If the feedback involves data errors, the engine will verify the customer's financial database or request the customer to upload corrected data.

[0104] Updates: Updates data points and related analysis in the report based on verified information.

[0105] 4) Report Adjustment

[0106] Report Restructuring: Adjust the structure and content of the assessment report based on client feedback. For example, add additional financial ratio analysis, or re-evaluate credit risk.

[0107] Customized analysis: For requirements of specific loan types or industries, the engine will use specific algorithms or models to perform more sophisticated analysis.

[0108] 5) Personalized content generation

[0109] Text generation: Use NLP to generate new text paragraphs to replace or add to reports, ensuring that the adjusted content is smooth, accurate and meets financial industry standards.

[0110] Visual presentation optimization: According to the changes in the report content, adjust the visual elements such as charts and tables accordingly to ensure the readability and professionalism of the report.

[0111] 6)Quality Control and Audit

[0112] Automated Review: The engine performs a built-in quality control process to ensure that all adjustments comply with the bank’s policies and risk management requirements.

[0113] Risk Assessment: Conduct a risk reassessment on the revised report to ensure that the adjustments do not introduce new risks.

[0114] 7) Generate personalized credit assessment report

[0115] Final report generation: After completing all adjustments, the engine generates a final personalized credit assessment report.

[0116] Report submission: Submit the personalized report to the corporate customer for final confirmation through the interactive user interface.

[0117] 8) Feedback loop

[0118] Client Confirmation: The client reviews the final report and provides confirmation or further feedback.

[0119] Iterative improvements: If there is further feedback, the engine may need to go back into the tuning process until the customer is fully satisfied.

[0120] Through the above process, the personalized adjustment engine ensures that the credit assessment report not only accurately reflects the customer's financial situation, but also fully considers the customer's special requirements and preferences, thereby providing a truly tailored loan service experience.

[0121] If the above method is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0122] Example 2

[0123] The present embodiment provides a personalized and interactive credit report generation device, comprising the following steps: a data acquisition module, a preliminary evaluation module and a personalized engine module, wherein the data acquisition module is used to acquire evaluation data of an evaluation object, wherein the evaluation data includes financial information and credit-related information; the preliminary evaluation module is used to generate a preliminary credit evaluation report based on the evaluation data using natural language processing technology; the personalized engine module is connected to an interactive user interface, and is used to receive feedback data transmitted by the interactive user interface, and determine whether the feedback data is report confirmation information. If not, the preliminary credit evaluation report is personalized adjusted based on the feedback data, and an updated credit evaluation report is generated, and feedback data is repeatedly received until the feedback data is report confirmation information, and the current credit evaluation report is output as the final credit evaluation report.

[0124] refer to Figure 4 As shown, the personalized and interactive credit report generation device is composed of a report generation system, which has a built-in intelligent language processing engine 1 (AI) and a personalized adjustment engine 2 (AE), and is connected to the client 4 (ClientSide), a database 3 (DB), etc. to provide the customer manager (CM) with a final confirmed credit assessment report. The report generation system and the database 3 (DB) are located in the bank internal system (Bank Internal System). The functions of the above components are as follows:

[0125] 1. Client Side

[0126] An interactive user interface (UI) is provided, which is the front-end platform for corporate customers to interact with the system. Corporate customers can log in, submit financial statements, view credit assessment reports, request report adjustments, and finally confirm reports through this interface.

[0127] 2. Bank Internal System

[0128] Used to provide banking data.

[0129] 3. Account Manager (CM)

[0130] Responsible for reviewing the final confirmed credit assessment report sent from the system of the present invention, and processing the loan request of the corporate customer based on the report.

[0131] 4. Report generation system

[0132] The core processing system is responsible for receiving data from the user interface, collecting the company's historical loans and product holdings from the database, using the intelligent language processing engine to generate and adjust the credit assessment report, and sending the final report to the account manager.

[0133] 5. Database 3 (DB)

[0134] Stores corporate financial data, historical loan records, deposit records, and other product holdings. The system will obtain the required data from here to perform credit assessment.

[0135] 6. Intelligent language processing engine 1 (AI)

[0136] It is used to analyze the data obtained from the database and generate a preliminary credit assessment report. The engine uses advanced natural language processing algorithms to understand the financial status and history of the enterprise.

[0137] 7. Personalization Engine 2 (AE)

[0138] Adjust the credit assessment report based on customer feedback on the user interface. This ensures that the report is closer to the actual needs of customers and also reflects personalized considerations for different loan types and varieties.

[0139] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0140] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0141] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0143] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0144] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A personalized and interactive credit report generation method, characterized in that, The following steps are involved: Obtaining assessment data of the assessment object, wherein the assessment data includes financial information and credit related information; Using natural language processing technology, generate a preliminary credit assessment report based on the assessment data; Receive feedback data, determine whether the feedback data is report confirmation information, and if not, adjust the preliminary credit assessment report based on the feedback data, generate an updated credit assessment report, repeatedly receive feedback data until the feedback data is report confirmation information, and output the current credit assessment report as the final credit assessment report.

2. The personalized and interactive credit report generation method according to claim 1, characterized in that: The financial information includes financial statements, historical loan records, deposit information and financial product holdings.

3. The personalized and interactive credit report generation method according to claim 1, characterized in that: The credit-related information includes credit rating and market reputation.

4. The personalized and interactive credit report generation method according to claim 1, characterized in that: A large language model is used to generate the preliminary credit assessment report.

5. The personalized and interactive credit report generation method according to claim 4, characterized in that: The large language model is pre-trained using large-scale financial field text data as training data, and the credit report data is used to fine-tune the model based on the pre-trained model.

6. The personalized and interactive credit report generation method according to claim 4, characterized in that: Based on a set update cycle, the large language model is retrained regularly.

7. The personalized and interactive credit report generation method according to claim 1, characterized in that: The feedback data includes error correction information, additional provision information and / or analysis result objection information.

8. The personalized and interactive credit report generation method according to claim 1, characterized in that: The personalized adjustment specifically includes: Utilize natural language processing technology to analyze the feedback data, understand the results based on the context, and determine whether there are data errors in the current credit assessment report. If so, verify the data and correct the errors; According to the analysis result of the feedback data, determining whether the report content needs to be adjusted, and if so, performing report adjustment processing based on the analysis result of the feedback data; Determine whether the adjusted report passes the quality review. If so, generate an updated credit assessment report.

9. The personalized and interactive credit report generation method according to claim 8, characterized in that: The quality review includes policy and risk review and risk reassessment.

10. A personalized and interactive credit report generating device, characterized in that, The following steps are involved: A data acquisition module, used to acquire the evaluation data of the evaluation object, wherein the evaluation data includes financial information and credit-related information; A preliminary assessment module, used to generate a preliminary credit assessment report based on the assessment data using natural language processing technology; The personalization engine module is connected to the interactive user interface and is used to receive feedback data transmitted by the interactive user interface, determine whether the feedback data is report confirmation information, and if not, adjust the preliminary credit assessment report based on the feedback data, generate an updated credit assessment report, repeatedly receive feedback data until the feedback data is report confirmation information, and output the current credit assessment report as the final credit assessment report.

Images