Intelligent financial rule configuration and financial report analysis method and system based on large model

Abstract

The invention relates to the technical field of large models, and provides an intelligent financial rule configuration and financial report analysis method and system based on a large model. The method comprises the following steps: acquiring a natural language instruction of a financial business demand, analyzing and mapping the natural language instruction to a corresponding rule template, and converting the natural language instruction into a structured rule field; based on the structured rule field, a large model is adopted, context information is introduced according to a financial report template and a specified dimension, and report content is generated; performing semantic understanding on report content to generate an analysis report; and abnormal values in the analysis report are identified, and correction suggestions and risk early warning are given. Through the natural language understanding ability of a large model, accurate mapping between natural language input and structured configuration of financial rules is realized, and the problems of high rule configuration threshold and high error rate in a traditional system are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of large models, in particular to a large model-based intelligent financial rule configuration and financial report analysis method and system. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute the prior art.

[0003] Currently, the financial rule configuration and financial report analysis in enterprise financial management are highly dependent on manual operation. On the one hand, financial rules (such as accounting engine configuration, transaction accounting rules, execution condition definition, etc.) are usually manually maintained by business personnel through an interface or a database, which has a high operation threshold, is prone to errors, and has a slow response to business changes. On the other hand, the data sources in the financial report generation process are complex and have various aspects, and manual interpretation is costly and time-consuming.

[0004] In the prior art, although some systems have attempted to introduce automated processing means, the rule configuration relies on professional skills, and non-technical personnel cannot directly participate, resulting in low configuration efficiency. Moreover, the financial report analysis relies on manual interpretation and lacks intelligent assistance, resulting in lagging analysis results. The rule conflict detection mechanism is not perfect, which can easily lead to logical confusion. The financial report is poorly explained, making it difficult for management to quickly understand the underlying reasons for the changes. It is impossible to achieve intelligent management of the entire process from rule input to financial report output. SUMMARY

[0005] To solve the technical problems in the background art, the present application provides a large model-based intelligent financial rule configuration and financial report analysis method and system. The present application realizes accurate mapping between natural language input and structured configuration of financial rules through the natural language understanding capability of large models, solving the problem of high rule configuration threshold and errors in traditional systems.

[0006] To achieve the above purpose, the present application adopts the following technical solutions: The first aspect of the present application provides a large model-based intelligent financial rule configuration and financial report analysis method.

[0007] A large model-based intelligent financial rule configuration and financial report analysis method, comprising: Obtaining natural language instructions of financial business requirements, parsing and mapping to corresponding rule templates, and converting natural language instructions to structured rule fields; Based on the structured rule fields, using a large model, generating report content according to the financial report template and the specified dimensions by introducing context information; performing semantic understanding on the report content to generate an analysis report; Identifying abnormal values in the analysis report and giving correction suggestions and risk warnings.

[0008] Further, in the process of converting natural language instructions into structured rule fields, the order of the rules is adjusted by dragging, the parameters are modified, and potential conflicts are fed back in real time.

[0009] Further, the financial report template comprises a metadata layer, a structure layer and a content layer, wherein the metadata layer comprises report number, name, type, applicable accounting standards and organization unit; the structure layer comprises a structure for describing the entire report and a content source for describing each cell; and the content layer comprises a cell-bound data source, a calculation formula and a conditional format.

[0010] Further, the context information is real-time context information obtained from a database or an interface.

[0011] Further, the semantic understanding of the report content comprises using NLP technology to perform semantic understanding on the report content, generating an abstract of financial report key indicators, identifying key elements in the financial report, judging the trend emotion in the financial report description, and identifying the causal relationship between different financial items.

[0012] Further, the construction and training process of the large model comprises constructing a financial term dictionary, constructing an entity recognition model in combination with accounting standards, performing fine-tuning on the entity recognition model, jointly analyzing the chart and text to obtain multi-modal fusion features, and guiding the entity recognition model to output financial-related language.

[0013] The second aspect of the present application provides an intelligent financial rule configuration and financial report analysis system based on a large model.

[0014] An intelligent financial rule configuration and financial report analysis system based on a large model comprises: A rule automatic recognition and configuration module configured to obtain natural language instructions of financial business requirements, parse and map to corresponding rule templates, and convert the natural language instructions into structured rule fields; A financial report automatic generation and analysis module configured to generate report content based on the structured rule fields, using a large model, according to the financial report template and the specified dimensions, and introducing context information; and perform semantic understanding on the report content to generate an analysis report; An abnormality detection and early warning module configured to identify abnormal values in the analysis report and give correction suggestions and risk warnings.

[0015] The third aspect of the present application provides a computer device, which comprises: A processor adapted to execute a computer program; The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the large model-based intelligent financial rule configuration and financial statement analysis method according to the first aspect.

[0016] The fourth aspect of the present application provides a computer readable storage medium storing a computer program, and the computer program is adapted to be loaded and executed by the processor to implement the steps of the large model-based intelligent financial rule configuration and financial statement analysis method according to the first aspect.

[0017] The fifth aspect of the present application provides a computer program product or a computer program.

[0018] The present application provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device execute the steps of the large model-based intelligent financial rule configuration and financial statement analysis method according to the first aspect.

[0019] Compared with the prior art, the present application has the following beneficial effects: The present application provides a large model-based intelligent financial rule configuration and financial statement analysis method and system. The method comprises the following steps: obtaining natural language instructions of financial business requirements, analyzing and mapping to corresponding rule templates, and converting the natural language instructions into structured rule fields; based on the structured rule fields, using a large model, generating report content according to a financial statement template and specified dimensions, introducing context information; performing semantic understanding on the report content to generate an analysis report; identifying abnormal values in the analysis report and giving correction suggestions and risk warnings. The present application improves the efficiency of financial rule configuration, and ordinary users can complete complex rule configuration without programming. The present application enhances the accuracy of financial statement generation, reduces human errors, and improves data consistency. The present application improves the depth and breadth of financial statement analysis, automatically generates multi-dimensional analysis and trend prediction, reduces compliance risks, automatically detects rule conflicts and financial statement abnormalities, and realizes full-link intelligent upgrading from financial rule configuration to financial statement analysis through deep integration of a large model and a financial system. BRIEF DESCRIPTION OF DRAWINGS

[0020] The drawings accompanying the specification of the present application form a part of the present application and serve to provide a further understanding of the present application, the illustrative embodiments of the present application and their descriptions serve to explain the present application, and do not constitute an improper limitation of the present application.

[0021] Figure 1 is a flowchart of the large model-based intelligent financial rule configuration and financial statement analysis method according to the embodiments of the present application. Figure 2 is a structural diagram of a large model-based intelligent financial rule configuration and financial report analysis system according to an embodiment of the present application; Figure 3 is an implementation flowchart of an algorithm corresponding to an anomaly detection and early warning module according to an embodiment of the present application; Figure 4 is a structural diagram of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The present application will be further described below in conjunction with the accompanying drawings and embodiments.

[0023] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0024] It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and furthermore, it should be understood that when the terms "comprise" and / or "include" are used in the specification, there is a presence of a feature, step, operation, device, component, and / or combinations thereof.

[0025] Figure 1 is a flowchart of a large model-based intelligent financial rule configuration and financial report analysis method according to an embodiment of the present application; refer to Figure 1 The method comprises: acquiring a natural language instruction of a financial business requirement, parsing and mapping to a corresponding rule template, and converting the natural language instruction to a structured rule field; based on the structured rule field, using a large model, introducing context information according to a financial report template and specified dimensions, generating report content; performing semantic understanding on the report content to generate an analysis report; identifying abnormal values in the analysis report and giving correction suggestions and risk warnings.

[0026] The present application is enhanced design based on the existing rule engine combined with LLM capability, mainly including the following innovative points: introducing LLM for automatic conversion from natural language to expression, dynamic context-aware execution framework, supporting multi-dimensional parameter injection (time, institution, currency, etc.), real-time error detection and feedback mechanism. The present application realizes root cause tracing and repair suggestion generation of financial report anomalies by integrating context analysis of business flow and financial data, significantly improving the efficiency of financial anomaly troubleshooting.

[0027] In some embodiments, in the process of converting natural language instructions into structured rule fields, the order of rules is adjusted by drag and drop, parameters are modified, and potential conflicts are fed back in real time.

[0028] In some embodiments, the financial report template includes a metadata layer, a structure layer, and a content layer, wherein the metadata layer includes report number, name, type, applicable accounting standards, and organization unit; the structure layer includes the structure for describing the entire report and the content source for describing each cell; and the content layer includes cell-bound data sources, calculation formulas, and conditional formats.

[0029] In some embodiments, the context information is real-time context information obtained from a database or an interface.

[0030] In some embodiments, the semantic understanding of the report content includes using NLP technology to perform semantic understanding on the report content, generating an abstract of financial report key indicators, identifying key elements in the financial report, judging the trend of emotions in the financial report description, and identifying the causal relationship between different financial items.

[0031] In some embodiments, the construction and training process of the large model includes constructing a financial term dictionary, constructing an entity recognition model in combination with accounting standards, fine-tuning the entity recognition model, jointly analyzing charts and text to obtain multi-modal fusion features, and guiding the entity recognition model to output financial-related language.

[0032] Figure 2 is a structural diagram of the intelligent financial rule configuration and financial report analysis system based on a large model according to an embodiment of the present application; refer to Figure 2 The present application is based on a large model (such as LLM, Prompt Engineering, fine-tuning, etc.) to construct an end-to-end intelligent financial system, which covers three core modules of automatic rule recognition and configuration, automatic generation and analysis of financial reports, and abnormality detection and early warning.

[0033] The automatic rule recognition and configuration module includes: (1) The user describes the financial business requirements in natural language (such as "configure the value-added tax price-tax separation rule for a certain product line"), and the system automatically analyzes and maps to the existing rule templates in the system; (2) Based on the knowledge graph and semantic matching, the natural language instructions are converted into structured rule fields, and consistency verification is performed in combination with historical rules; (3) Support for multiple language inputs, automatic translation into the system default language (such as Chinese / English), and ensure the adaptation of international accounting standards (such as IFRS, GAAP); The application realizes multi-language support and automatic conversion of different accounting standards (such as IFRS, GAAP) for multinational enterprise financial reports through multi-language large model and internationalization adaptation technology, and improves the global applicability of the system.

[0034] (4) A visual rule editor is provided, and users can drag and adjust the order of rules, modify parameters, and real-time feedback potential conflicts (such as subject code duplication, start and end time overlap, etc.).

[0035] The application realizes semantic association between financial rules, accounting subjects, business processes, and financial report structure by constructing a multi-modal knowledge graph, and improves the understanding ability of the system for complex business scenarios.

[0036] In some embodiments, the financial report automatic generation and analysis module comprises: (1) The large model automatically generates report content according to the financial report template and the specified dimensions (such as account set, organization, currency, and period); The financial report template is one of the core inputs of the entire system, which determines the structure, content and presentation of the final report. Its design includes the following aspects: 1) Template structure design ① Metadata layer (Metadata Layer), including: report number, name, type (balance sheet, income statement, cash flow statement, etc.), applicable accounting standards (IFRS / GAAP / Chinese accounting standards), organization unit (company, department, group) etc.; ② Structure layer (Structure Layer), row and column definition: using ReportTemplateInfo and ReportcellInfo objects. ReportTemplateInfo: describes the structure of the entire report (such as number of rows, number of columns, title, etc.); ReportcellInfo: describes the content source of each cell (formula, fixed value, reference to other reports, etc.).

[0037] ③ Content layer (Content Layer): cell binding data source (such as subject code, auxiliary accounting field), dynamic calculation formula (supporting scripting languages such as Groovy, JavaScript or DSL), conditional format (such as negative number red display).

[0038] 2) Template storage and management Storage form: JSON, XML or database table structure; Version control: support multi-version management, convenient for audit traceability; Visual editor: provides a drag-and-drop interface, supports preview and verification functions.

[0039] (2) Dynamic calculation combined with contextual information (such as exchange rate, tax rate, subject balance) to avoid hard-coded formula errors; This process mainly realizes intelligent dynamic calculation of financial data through the combination of large models (LLM) and business logic. The specific process is as follows.

[0040] 1) Context extraction: The system obtains real-time contextual information from the database or interface, such as exchange rate (based on transaction date and currency), tax rate (according to region, industry, tax type, etc.), subject balance (general ledger subject, auxiliary accounting item), time dimension (accounting period, year).

[0041] 2) Rule analysis and execution engine: Convert natural language described financial rules into structured expressions; use rule engine (such as Drools) to convert rules into executable logic. Inject context variables during execution for dynamic calculation.

[0042] 3) Large model assisted reasoning: LLM is used to understand the implicit conditions in complex business logic (such as "monthly transfer of undistributed profits"), and generate corresponding calculation logic. For ambiguous or incomplete rules, LLM can perform semantic completion and prompt the user for confirmation.

[0043] Through the reasoning ability of large models and the context perception mechanism, the invention realizes the dynamic calculation and anomaly detection of financial data, avoiding the errors caused by traditional hard-coded formulas.

[0044] (3) Use NLP technology to understand the semantics of financial report content and automatically generate a text version of the analysis report (such as income growth analysis, cost structure change, and reasons for profit fluctuation): This process is not just a simple call to NLP technology, but a deep optimization and improvement combined with financial field knowledge, mainly including the following aspects: 1) Application of general NLP technology Text summary: Use a Transformer-based model (such as BERT, T5) to generate a summary of key financial indicators; Keyword extraction: Identify key elements in the financial report (such as revenue, net profit, cost changes, etc.); Sentiment analysis: Determine the trend sentiment (growth, decline, stability) in the financial report description; Relationship extraction: Identify the causal relationship between different financial items (such as sales growth leading to profit increase); The present application realizes automatic interpretation of financial report content and generation of analysis report through NLP generation technology, greatly reduces the time cost of manual writing of financial analysis report, enhances the decision support capability, and assists the management layer in making quick decisions through intelligent analysis report.

[0045] 2) Improvement and optimization of the financial field ① Construction of field dictionary: construct a financial term dictionary (such as "depreciation", "amortization", "deferred income tax", etc.), and construct an entity recognition model (NER) combined with accounting standards.

[0046] ② Fine-tuning model training: fine-tune the pre-trained model (such as Chinese-BERT-wwm-ext) based on the data set derived from historical financial reports, analyst reports, and announcement documents.

[0047] ③ Multi-modal fusion: joint analysis of charts and text (such as consistency check of bar charts and corresponding text description), and use of chart OCR recognition and text semantic matching.

[0048] ④ Controllable generation: use Prompt Engineering to guide the model to output language that conforms to the financial style, and control the accuracy and compliance of the generated content (avoid subjective speculation).

[0049] (4) Support multi-dimensional drilling analysis (such as splitting data by product, customer, and region dimensions), and present key indicator trends through charts and text combination.

[0050] Figure 3 is the implementation flowchart of the abnormality detection and early warning module corresponding to the algorithm shown in the embodiment of the present application; refer to Figure 3 , the abnormality detection and early warning module is configured to realize: (1) In the rule configuration stage, the large model automatically detects potential conflicts (such as duplicate amount fields, disabled subjects, cross-system configuration errors, etc.); (2) After the financial report is generated, combined with historical data and industry benchmarks, identify outliers (such as sudden increase in income, unreasonable decrease in expenses, etc.); (3) Automatically associate related business flow (such as original business flow, journal flow, etc.) for root cause analysis, and give correction suggestions; (4) Pre-set risk level and early warning strategy, support multiple ways of pushing early warning information such as email, SMS, and system notification.

[0051] The present application realizes intelligent error correction and optimization suggestions in the rule configuration process through reinforcement learning and rule conflict detection algorithm, improves the system stability and user experience.

[0052] This invention enhances the existing rule engine by combining LLM capabilities. Its main innovations include the introduction of LLM for automatic conversion of natural language into expressions, a dynamic context-aware execution framework, support for multi-dimensional parameter injection (time, organization, currency, etc.), and a real-time error detection and feedback mechanism.

[0053] For example: A company needs to configure VAT price-tax separation rules for its "smart home product line." The natural language instructions are: For the sales revenue of the smart home product line, price-tax separation should be performed at a 13% VAT rate. The tax amount after separation should be recorded in the "Taxes Payable - VAT Payable (Output Tax)" account. The price-tax separation results should be summarized in the financial statements on a monthly basis.

[0054] 1. Natural Language Parsing and Structured Mapping 1. Key element extraction Business Target: Sales revenue of smart home product line Rule type: VAT price-tax separation Tax rate: 13% Accounting subject: Taxes payable - VAT payable (output tax) Time dimension: Monthly summary 2. Convert to structured expression The natural language understanding (NLU) capabilities of the large model break down the instructions into structured fields that can be recognized by the system (using JSON format as an example): { "rule_name": "Smart Home Product Line - VAT Price and Tax Separation", "business_object": "Sales Revenue", "product_line": "Smart Home", "tax_rate": 0.13, "tax_account": "Taxes payable - VAT payable (output tax)", "time_dimension": "monthly", "calculation_logic": "Price-tax separation" } 2. Rule Engine (Drools) Executes Logical Transformations 1. Rule engine rule definition Use Drools to convert structured expressions into executable business rules (Java pseudocode example): rule "Smart Home Product Line - VAT Price and Tax Separation Rules" when / / Match transaction data where business object is "Sales Revenue" and product line is "Smart Home" Transaction( businessObject == "Sales Revenue", productLine == "Smart Home", $amount: amount, $transactionDate: transactionDate ) then / / Calculate tax amount and net amount (inject tax rate, accounting code, etc. context variables) double taxRate = 0.13; / / Get tax rate from structured expression double taxAmount = $amount × taxRate; / / Tax amount = taxable amount × tax rate double netAmount = $amount - taxAmount; / / Net amount = taxable amount - tax amount / / Create accounting entries (inject code, etc. context variables) create(new AccountingEntry( "Tax Payable - VAT (Output Tax)", taxAmount, $transactionDate, "Tax Separation - Smart Home Product Line" )); / / Accumulate results by month (inject time dimension context variables) accumulate(Transaction, filter(productLine == "Smart Home" && month($transactionDate) == currentMonth), sum(amount) -> $monthlyTotal ); end 2. Context variable injection Real-time data: Get current month (currentMonth), transaction date ($transactionDate), etc. dynamic information from the database.

[0055] Business parameters: Extract tax rate (0.13) and accounting subject (Tax payable - VAT (output tax)) from structured expressions.

[0056] Dynamic calculation: Calculate tax amount (taxAmount) and net amount (netAmount) based on transaction amount ($amount) and tax rate in real time.

[0057] Three, dynamic calculation and result output 1. Execution process Data input: The business system submits a sales revenue data of smart home product line (such as tax-inclusive amount 11300 yuan, transaction date June 5, 2025).

[0058] Rule matching: The Drools engine detects that the data meets the "smart home product line - VAT price and tax separation rule" condition.

[0059] Dynamic calculation: Tax amount = 11300 × 13% = 1300 yuan Net amount = 11300 - 1300 = 10000 yuan Result output: Generate accounting entries: Debit: Bank deposit 11300 Credit: Main business income - smart home 10000 Credit: Tax payable - tax payable (output tax) 1300 Monthly summary data: The total sales revenue of smart home product line in June is 1 million yuan, and the total tax amount is 13 million yuan.

[0060] 2. Financial report association Net amount is included in the "main business income" subject of the profit statement; Tax amount is included in the "tax payable" subject of the balance sheet; Monthly summary data is output to the financial statements through the financial report automatic generation module, and triggers NLP analysis to generate a text explanation (such as "smart home product line VAT output tax increased by 5% in June compared with the same period last year, mainly due to the increase in sales revenue").

[0061] The present application not only solves the many pain points of the prior art in financial rule configuration and financial report analysis, but also provides a solid technical foundation for the future development of financial intelligence.

[0062] See Figure 4A structural diagram of a computer device is shown, the computer device including a processor, a communication interface, and a computer readable storage medium. The processor, the communication interface, and the computer readable storage medium are connected through a bus or other means. The communication interface is configured to receive and send data. The computer readable storage medium can be stored in the memory of the computer device, and the computer readable storage medium is configured to store a computer program including program instructions, and the processor is configured to execute the program instructions stored in the computer readable storage medium. The processor (or CPU) is the computing core and control core of the computer device, and is adapted to implement one or more instructions, and is specifically adapted to load and execute one or more instructions to implement the corresponding steps in the embodiment of the intelligent financial rule configuration and financial report analysis method based on a large model.

[0063] The embodiment provides a computer readable storage medium (Memory), which is a memory device in a computer device and is used for storing programs and data. It can be understood that the computer readable storage medium herein can include an internal storage medium in the computer device, and of course can include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space, and the storage space stores a processing system of the computer device. In addition, one or more instructions adapted to be loaded and executed by the processor are also stored in the storage space, and the instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory; optionally, the computer readable storage medium can also be at least one computer readable storage medium located away from the aforementioned processor.

[0064] In one embodiment, the computer readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer readable storage medium to implement the corresponding steps in the above-mentioned embodiment of the intelligent financial rule configuration and financial report analysis method based on a large model.

[0065] The embodiment provides a computer program product or computer program, which includes computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the corresponding steps in the above-mentioned embodiment of the intelligent financial rule configuration and financial report analysis method based on a large model.

[0066] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, 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 magnetic disk storage and optical storage, etc.) containing computer-usable program code.

[0067] 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 flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, 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 flowcharts and / or block diagrams. 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.

[0068] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0070] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0071] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. Intelligent financial rule configuration and financial report analysis method based on large models, characterized by: include: Obtain natural language instructions for financial business needs, parse and map them to corresponding rule templates, and convert natural language instructions into structured rule fields; Based on structured rule fields, a large model is used to introduce contextual information and generate report content according to the financial report template and specified dimensions; Conduct semantic understanding of report content and generate analysis reports; Identify outliers in analysis reports and provide correction suggestions and risk warnings.

2. The method for configuring intelligent financial rules and analyzing financial reports based on a large model according to claim 1, characterized in that: In the process of converting natural language instructions into structured rule fields, it is supported to adjust the rule order and modify parameters by dragging and dropping, and provide real-time feedback on potential conflicts.

3. The method for configuring intelligent financial rules and analyzing financial reports based on a large model according to claim 1, characterized in that: The financial report template includes: a metadata layer, a structure layer and a content layer, wherein the metadata layer includes the report number, name, type, applicable accounting standards and the organizational unit to which it belongs; the structure layer includes: a layer for describing the structure of the entire report and a layer for describing the content source of each cell; the content layer includes: a data source bound to a cell, a calculation formula and conditional formatting.

4. The method for configuring intelligent financial rules and analyzing financial reports based on a large model according to claim 1, characterized in that: The context information is real-time context information obtained from a database or an interface.

5. The method for configuring intelligent financial rules and analyzing financial reports based on a large model according to claim 1 is characterized in that: The semantic understanding of the report content includes: using NLP technology to perform semantic understanding of the report content, generating summaries of key financial report indicators, identifying key elements in the financial report, judging trend emotions in the financial report description, and identifying causal relationships between different financial items.

6. The method for configuring intelligent financial rules and analyzing financial statements based on a large model according to claim 1, characterized in that: The construction and training process of the large model includes: building a dictionary of financial terms, combining accounting standards, and building an entity recognition model; fine-tuning the entity recognition model, jointly analyzing charts and text to obtain multimodal fusion features; guiding the entity recognition model to output finance-related language.

7. Intelligent financial rule configuration and financial report analysis system based on large models, characterized by: include: The rule automatic identification and configuration module is configured to: obtain natural language instructions of financial business requirements, parse and map them to corresponding rule templates, and convert natural language instructions into structured rule fields; The automatic financial report generation and analysis module is configured to: generate report content based on structured rule fields, using a large model, and introducing contextual information according to the financial report template and specified dimensions; Conduct semantic understanding of report content and generate analysis reports; The anomaly detection and warning module is configured to: identify anomalies in analysis reports and provide correction suggestions and risk warnings.

8. A computer device, characterized in that: a processor adapted to execute a computer program; A computer-readable storage medium having a computer program stored therein, wherein the computer program, when executed by the processor, implements the steps of the large-model-based intelligent financial rule configuration and financial report analysis method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which is suitable for being loaded by a processor and executing the steps in the large-model-based intelligent financial rule configuration and financial report analysis method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the steps of the large model-based intelligent financial rule configuration and financial report analysis method according to any one of claims 1 to 6.

Images