An intelligent financial analysis method and device, electronic equipment and storage medium

Abstract

This application provides an intelligent financial analysis method, apparatus, electronic device, and storage medium, comprising: determining financial indicators corresponding to each financial analysis strategy based on the financial data to be analyzed; determining prompt words corresponding to each financial analysis strategy based on prompt word engineering and financial analysis templates; creating a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompt words, and a pre-trained intelligent financial analysis model corresponding to each financial analysis strategy; creating a financial analysis knowledge graph; and generating a financial analysis report corresponding to each financial analysis strategy using the target intelligent financial analysis model, financial analysis template, financial indicators, and financial analysis knowledge graph. This application generates financial analysis reports by creating intelligent financial analysis models and financial analysis knowledge graphs, reducing manual creation costs and error rates, and improving the efficiency, accuracy, and reliability of financial report creation.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to an intelligent financial analysis method, device, electronic device, and storage medium. Background Technology

[0002] Traditional financial management reports are based on financial information and involve the disclosure of data from multiple dimensions and indicators. This involves streamlining common data tables such as the balance sheet, income statement, and cash flow statement, and combining them with the company's actual situation to achieve multi-dimensional analysis and indicator generation. Prior to this, the business architecture of financial analysis needs to be defined. After abstracting and aggregating data indicators such as sales revenue, sales volume, cost, inventory, and funds within the business framework of financial analysis, the needs of report users are converted into dimensions and indicators. The system's application architecture is then built based on these dimensions and indicators. This process significantly increases the workload of financial personnel, requiring complex spreadsheet designs and linked formulas to achieve indicator results. Not only is finding indicator data complex, but it also increases the risk of errors in manual data retrieval, reducing the accuracy and reliability of the final financial report. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide at least one intelligent financial analysis method, device, electronic device and storage medium, which generates financial analysis reports by creating intelligent financial analysis models and financial analysis knowledge graphs, thereby reducing manual creation costs and error rates, and improving the efficiency, accuracy and reliability of financial report creation.

[0004] This application mainly includes the following aspects:

[0005] In a first aspect, embodiments of this application provide an intelligent financial analysis method, the method comprising: collecting financial data to be analyzed corresponding to a target data source; determining multiple financial analysis strategies based on different user financial analysis needs; determining financial indicators corresponding to each financial analysis strategy based on the financial data to be analyzed; determining prompt words corresponding to each financial analysis strategy based on prompt word engineering and financial analysis templates; creating a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompt words, and pre-trained intelligent financial analysis model corresponding to each financial analysis strategy; and generating a financial analysis report corresponding to each financial analysis strategy using the target intelligent financial analysis model, financial analysis template, financial indicators, and financial analysis knowledge graph.

[0006] In one possible implementation, the target data source includes multiple business system interfaces, wherein the financial data to be analyzed is determined by: automatically identifying and collecting corresponding financial data from each business system interface using robotic process automation (Robotic Process Automation) technology; cleaning the financial data to obtain cleaned financial data, which includes at least deduplication, error correction, and missing value imputation; validating the cleaned financial data, and if the cleaned financial data passes the data validation, standardizing the cleaned financial data to obtain the financial data to be analyzed, which includes at least consistency verification and reasonableness verification; and generating an error data alarm if the cleaned financial data fails the data validation, which includes the error type and error data location.

[0007] In one possible implementation, the financial analysis template includes multiple financial analysis items, wherein prompts corresponding to each financial analysis strategy are created by: identifying target financial analysis items in the financial analysis template, the analysis content corresponding to the target financial analysis items being determined by an intelligent financial analysis model; generating content prompts corresponding to each target financial analysis item using prompt engineering for each target financial analysis item; and forming prompts corresponding to the financial analysis strategy from the content prompts corresponding to each target financial analysis item.

[0008] In one possible implementation, the target intelligent financial analysis model is created as follows: For each financial analysis strategy, the following processing is performed: From the corresponding financial indicators, determine multiple financial analysis indicators corresponding to each target financial analysis item in the financial analysis template corresponding to the financial analysis strategy; Input the multiple financial analysis indicators and content prompts corresponding to each target financial analysis item into the pre-trained intelligent financial analysis model to obtain the analysis content corresponding to each target financial analysis item; Fine-tune the intelligent financial analysis model using the analysis content corresponding to each target financial analysis item of each financial analysis strategy to obtain the target intelligent financial analysis model.

[0009] In one possible implementation, a financial analysis knowledge graph is created by extracting financial knowledge and business logic from the financial data to be analyzed, industry reports, and regulatory documents; and organizing the financial knowledge and business logic in the form of a graph to form a financial knowledge graph.

[0010] In one possible implementation, a financial analysis report corresponding to each financial analysis strategy is generated as follows: The financial analysis indicators corresponding to each financial analysis item in the financial analysis template corresponding to the financial analysis strategy are determined; for each financial analysis item, the following processing is performed: the generation method of the analysis content corresponding to the financial analysis item is determined; if the analysis content corresponding to the financial analysis item is generated by a large model, the content prompts and financial analysis indicators corresponding to the financial analysis item are input into the target intelligent financial analysis model to obtain the analysis content corresponding to the financial analysis item; if the analysis content corresponding to the financial analysis item is generated by the aggregation of financial indicators, the multiple financial analysis indicators corresponding to the financial analysis item are directly determined as the analysis content corresponding to the financial analysis item; if the analysis content corresponding to the financial analysis item is formed by a financial analysis knowledge graph, the corresponding financial knowledge is extracted from the financial analysis knowledge graph based on the financial analysis indicators corresponding to the financial analysis item, and the financial knowledge is determined as the analysis content corresponding to the financial analysis item; the analysis content corresponding to each financial analysis item is filled into the corresponding position of the financial analysis item; and the filled financial analysis items form the financial analysis report corresponding to the financial analysis strategy.

[0011] In one possible implementation, before filling the corresponding position of each financial analysis item with the analysis content, the method further includes: for each financial analysis item, performing the following processing: determining the target content format corresponding to the financial analysis item, the content format including at least text description, data table and chart; converting the analysis content corresponding to the financial analysis item into the target content format.

[0012] In one possible implementation, the financial analysis report corresponding to each financial analysis strategy is stored in a preset storage location in the form of a vector. The method further includes: receiving a financial data query request input from a requesting end; parsing the financial data query request to determine the query type, which may include financial indicator query, financial report query, and financial knowledge query; if the financial data query request is a financial indicator query, outputting multiple target financial analysis indicators corresponding to the financial indicator query as response results and returning them to the requesting end; if the financial data query request is a financial report query, determining the target financial report type indicated by the financial data query request, retrieving the financial analysis report corresponding to the target financial report type from the preset storage location as response results and returning it to the requesting end; if the financial data query request is a financial knowledge query, extracting the query keywords corresponding to the financial data query, extracting the financial knowledge corresponding to the query keywords from the financial analysis knowledge graph as response results and returning it to the requesting end.

[0013] In one possible implementation, the method further includes: acquiring historical financial data generated over a preset historical time period; constructing a macroeconomic forecasting model and determining the forecast results based on the historical financial data and macroeconomic indicators; and generating a decision support report based on the forecast results, the trend of financial data changes, and a preset decision algorithm.

[0014] In one possible implementation, the method further includes: receiving a financial data display request sent by a requesting end; parsing the financial data display request to determine the query keywords, analysis dimensions, and target display format corresponding to the target financial data; obtaining the target financial data according to the query keywords; performing dimensional analysis on the target financial data according to the analysis dimensions to obtain the target sub-financial data corresponding to each analysis dimension; converting the target sub-financial data into a target display type; and feeding back the converted target sub-financial data to the requesting end.

[0015] This application provides an intelligent financial analysis method, apparatus, electronic device, and storage medium, comprising: determining multiple financial analysis strategies based on different user financial analysis needs; determining financial indicators corresponding to each financial analysis strategy based on the financial data to be analyzed; determining prompt words corresponding to each financial analysis strategy based on prompt word engineering and financial analysis templates; creating a target intelligent financial analysis model based on the financial analysis template, financial indicators, and prompt words corresponding to each financial analysis strategy; creating a financial analysis knowledge graph; and generating a financial analysis report corresponding to each financial analysis strategy using the target intelligent financial analysis model, financial analysis template, financial indicators, and financial analysis knowledge graph, and saving the financial analysis report corresponding to each financial analysis strategy to a preset storage location. This application generates financial analysis reports by creating intelligent financial analysis models and financial analysis knowledge graphs, reducing manual creation costs and error rates, and improving the efficiency, accuracy, and reliability of financial report creation.

[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This document illustrates one of the flowcharts of an intelligent financial analysis method provided in an embodiment of this application;

[0019] Figure 2This illustrates a second flowchart of an intelligent financial analysis method provided in an embodiment of this application;

[0020] Figure 3 This paper illustrates a functional block diagram of an intelligent financial analysis device provided in an embodiment of this application.

[0021] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0023] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0024] Traditional financial management reports are based on financial information and involve the disclosure of data from multiple dimensions and indicators. This involves streamlining common data tables such as the balance sheet, income statement, and cash flow statement, and combining them with the company's actual situation to achieve multi-dimensional analysis and indicator generation. Prior to this, the business architecture of financial analysis needs to be defined. After abstracting and aggregating data indicators such as sales revenue, sales volume, cost, inventory, and funds within the business framework of financial analysis, the needs of report users are converted into dimensions and indicators. The system's application architecture is then built based on these dimensions and indicators. This process significantly increases the workload of financial personnel, requiring complex spreadsheet designs and linked formulas to achieve indicator results. Not only is finding indicator data complex, but it also increases the risk of errors in manual data retrieval, reducing the accuracy and reliability of the final financial report.

[0025] Based on this, this application provides an intelligent financial analysis method that generates financial analysis reports by creating intelligent financial analysis models and financial analysis knowledge graphs, reducing manual creation costs and error rates, and improving the efficiency, accuracy, and reliability of financial report creation, as detailed below:

[0026] Please see Figure 1 , Figure 1 This document illustrates one of the flowcharts for an intelligent financial analysis method provided in an embodiment of this application. For example... Figure 1 As shown, the method provided in this application embodiment includes the following steps:

[0027] S100. Collect the financial data to be analyzed corresponding to the target data source.

[0028] S200: Based on the different financial analysis needs of users, determine multiple financial analysis strategies.

[0029] S300. Based on the financial data to be analyzed, determine the financial indicators corresponding to each financial analysis strategy.

[0030] S400: Determine the prompt words corresponding to each financial analysis strategy based on prompt word engineering and financial analysis templates.

[0031] S500: Create a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompts, and pre-trained intelligent financial analysis model corresponding to each financial analysis strategy.

[0032] S600 utilizes the target intelligent financial analysis model, financial analysis templates, financial indicators, and financial analysis knowledge graph to generate financial analysis reports corresponding to each financial analysis strategy.

[0033] In specific implementation, in steps S100 to S700 provided in this application, the pre-trained intelligent financial analysis model is the qwen-finance large model.

[0034] Prompt engineering, also known as contextual prompting, refers to refining prompts using structured text and other methods to guide large models to output the desired results. Prompt engineering allows large models to complete different types of tasks without updating model weights.

[0035] This application combines prompt word engineering, financial analysis templates, and the qwen-finance big model. Utilizing the AI ​​capabilities provided by the qwen-finance big model, and guided by prompt word engineering, financial analysis templates, and financial analysis knowledge graphs, it automatically analyzes and mines financial data to quickly and accurately generate financial reports that match the company's current operating situation, providing strong support for the company's steady development.

[0036] The process of automating the generation of financial analysis reports in this application significantly reduces enterprises' reliance on human resources, enabling them to drastically reduce their investment in repetitive labor, effectively lowering their operating costs, and significantly improving resource utilization efficiency. Furthermore, this application generates financial analysis reports that quickly and accurately reflect the current operating conditions of enterprises, providing strong support for their steady development.

[0037] In a preferred embodiment, the target data source includes multiple business system interfaces, and step S100 includes:

[0038] Using Robotic Process Automation (RPA) technology, corresponding financial data is automatically identified and collected from each business system interface. The financial data is then cleaned to obtain cleaned financial data, which includes at least deduplication, error correction, and missing value imputation. The cleaned financial data is then validated. If the cleaned financial data passes validation, it is standardized to obtain the financial data to be analyzed. Data validation includes at least consistency and reasonableness verification. If the cleaned financial data fails validation, an error data alarm is generated, which includes the error type and error location.

[0039] Financial data includes, but is not limited to, at least one of the following: financial statements, transaction records, and market data. In this application, financial data is first cleaned to ensure its quality. Then, the cleaned financial data is validated to detect anomalies. This validation process provides real-time feedback when anomalies or errors occur, helping users to identify and correct errors early and preventing them from propagating to the final financial analysis report. This improves the accuracy of the final report. Furthermore, the standardization of financial data converts data from different sources and formats into a unified format, facilitating subsequent processing and analysis.

[0040] In a preferred embodiment, in step S200, multiple financial analysis strategies can be constructed by combining different financial analysis needs of users with the experience of business experts. These financial analysis strategies essentially indicate financial analysis rules corresponding to the financial analysis needs. The multiple financial analysis strategies include, but are not limited to, at least one of the following:

[0041] Financial performance indicator analysis strategies, profitability analysis strategies, operational efficiency analysis strategies, debt repayment efficiency analysis strategies, project development prospect analysis strategies, development capability analysis strategies, DuPont analysis strategies, project economic benefit evaluation analysis strategies, capital structure analysis strategies, return on investment capability analysis strategies, financial ratio analysis strategies, and accounts receivable and inventory control analysis strategies.

[0042] In step S300, based on the financial analysis strategy, financial indicators corresponding to the financial analysis strategy are determined from the financial data to be analyzed. Specifically, the financial indicators are divided into different dimension levels, and different dimension levels correspond to different financial indicators. The financial indicators include, but are not limited to, at least one of the following: total assets, total liabilities, total liabilities and shareholders' equity, and current assets.

[0043] In a preferred embodiment, the financial analysis template includes multiple financial analysis items, and the prompts corresponding to the financial analysis strategies include content prompts corresponding to each financial analysis item. Step S400 includes:

[0044] The target financial analysis items in the financial analysis template are determined. The analysis content corresponding to the target financial analysis items is determined by the intelligent financial analysis model. For each target financial analysis item, the prompt words are generated using prompt word engineering. The prompt words corresponding to each target financial analysis item form the prompt words corresponding to the financial analysis strategy.

[0045] Specifically, each financial analysis item in the financial analysis template can be edited and determined according to the user's actual analysis needs. In this application, it is necessary to pre-import the financial analysis template corresponding to each financial analysis strategy or use the online editing function of the financial analysis template provided by the financial statement analysis system. Combined with the actual analysis needs, the financial analysis template corresponding to the financial analysis strategy is designed online. The generation method of the analysis content corresponding to each financial analysis item is pre-specified. Content prompts are used to guide the intelligent financial analysis model to output the analysis content corresponding to the financial analysis item. However, in the process of creating the corresponding financial analysis report using the pre-trained intelligent financial analysis model, it is possible that not all the analysis content corresponding to all financial analysis items in the financial analysis template needs to be determined by the intelligent financial analysis model. For example, some financial analysis items may simply be an integration of financial analysis indicators. Therefore, when creating content prompts, only the corresponding content prompts need to be determined for the financial analysis items that require intelligent financial analysis to obtain the corresponding analysis content.

[0046] In this application, users can customize the format, content, and level of detail of the financial analysis template according to their needs, and the system will generate a customized financial analysis report according to the user's requirements.

[0047] In a preferred embodiment, step S500 includes:

[0048] For each financial analysis strategy, the following processing is performed: From the corresponding financial indicators, determine multiple financial analysis indicators corresponding to each target financial analysis item in the financial analysis template corresponding to the financial analysis strategy; input the multiple financial analysis indicators and content prompts corresponding to each target financial analysis item into the pre-trained intelligent financial analysis model to obtain the analysis content corresponding to each target financial analysis item; and fine-tune the intelligent financial analysis model using the analysis content corresponding to each target financial analysis item of each financial analysis strategy to obtain the target intelligent financial analysis model.

[0049] In one example, this application employs RLHF (Reinforcement Learning from Human Feedback) to pre-train the intelligent financial analysis model. This training method minimizes invalid, distorted, or biased outputs generated by the pre-trained intelligent financial analysis model. Then, this application uses multiple financial analysis indicators and content prompts corresponding to each target financial analysis item under each financial analysis strategy to fine-tune the pre-trained intelligent financial analysis model until the error between the analysis content corresponding to each target financial analysis item under each financial analysis strategy output by the intelligent financial analysis model and the standard analysis content is within a preset error range. This yields the fine-tuned target intelligent financial analysis model, i.e., the target intelligent financial analysis model. The process of fine-tuning the intelligent financial analysis model actually involves adjusting the content prompts corresponding to the target financial analysis items until the intelligent financial analysis model can obtain the optimal analysis content under the guidance of the content prompts.

[0050] This application allows for fine-tuning of pre-trained intelligent financial analysis models based on the specific needs of different industries, optimizing the models to adapt to industry characteristics and improving the accuracy and reliability of financial analysis.

[0051] Before determining the target intelligent financial analysis model, it is necessary to evaluate the performance of the fine-tuned intelligent financial analysis model through validation and test sets to ensure the accuracy and generalization ability of the final target intelligent financial analysis model.

[0052] In this application, by automating the data collection, processing, and analysis process, intelligent financial reporting reduces human intervention during generation, enabling enterprises to obtain key business information more quickly and accelerating decision-making processes. Simultaneously, based on big data and machine learning algorithms, AI can deeply analyze complex datasets, discover hidden patterns and trends, and provide management with more accurate and comprehensive business insights. Through data-driven decision-making supported by deep analysis, enterprises can reduce blind spots and subjectivity, improving the scientific rigor and accuracy of their decisions.

[0053] In a preferred embodiment, the method provided in this application further includes:

[0054] Financial knowledge and business logic are extracted from the financial data to be analyzed, industry reports, and regulatory documents, and then organized in the form of a graph to form a financial knowledge graph.

[0055] Specifically, natural language processing technology is used to collect financial data to be analyzed, as well as financial information such as industry reports and regulatory documents related to the financial field. Key information is automatically extracted to form structured and semi-structured data. This data includes, but is not limited to, the company's financial data, financial knowledge and related knowledge points, relationships and business logic. Then, this data is normalized, categorized and semantically processed, that is, the data is converted into a highly understandable language form. Then, using technologies such as ontolgy and graph, it is stored in a graph database to construct a financial analysis knowledge graph. This ensures that the company obtains the compliance of the business processes corresponding to the financial data to be analyzed. Moreover, this application regularly updates the financial analysis knowledge graph so that it can reflect the latest financial regulations and business practices.

[0056] By integrating information such as financial indicators, business processes, and regulatory compliance, a financial knowledge graph is constructed to provide rich background knowledge for intelligent analysis.

[0057] In a preferred embodiment, in step S700, for each financial analysis strategy, the following is performed:

[0058] Determine the financial analysis indicators corresponding to each financial analysis item in the financial analysis template for this financial analysis strategy, and perform the following processing for each financial analysis item:

[0059] Determine how the analysis content for this financial analysis item will be generated.

[0060] If the analysis content corresponding to the financial analysis item is generated by the large model, then input the content prompts and financial analysis indicators corresponding to the financial analysis item into the target intelligent financial analysis model to obtain the analysis content corresponding to the financial analysis item.

[0061] If the analysis content corresponding to a financial analysis item is generated by aggregating financial indicators, then the multiple financial analysis indicators corresponding to that financial analysis item will be directly determined as the analysis content corresponding to that financial analysis item.

[0062] If the analysis content corresponding to a financial analysis item is formed by a financial analysis knowledge graph, then the business process corresponding to each financial analysis indicator is extracted from the financial analysis knowledge graph, and the business process is determined as the analysis content corresponding to that financial analysis item.

[0063] Fill in the corresponding analysis content for each financial analysis item into the corresponding position for that financial analysis item.

[0064] Each completed financial analysis item forms the corresponding financial analysis report for that financial analysis strategy.

[0065] In one specific embodiment, if the analysis content corresponding to the financial analysis item is specified to be generated by the target intelligent financial analysis model, then the content prompt words and financial analysis indicators corresponding to the financial analysis item are jointly input into the target intelligent financial analysis model, the data output by the target intelligent financial analysis model is used as the analysis content corresponding to the financial analysis item, and the analysis content is mapped to the analysis content filling area corresponding to the financial analysis item.

[0066] If the analysis content corresponding to the financial analysis item is specified to be generated by the aggregation of financial indicators, then the multiple financial analysis indicators corresponding to the financial analysis item will be directly used as the analysis content and mapped to the corresponding analysis content filling area.

[0067] If the analysis content corresponding to the financial analysis item is specified as financial knowledge, then the corresponding financial knowledge is extracted from the financial analysis knowledge graph based on the financial analysis indicators corresponding to the financial analysis item as the analysis content corresponding to the financial analysis item, and the corresponding analysis content is mapped to the corresponding analysis content filling area.

[0068] This application utilizes a finely tuned large model and a financial analysis knowledge graph for in-depth analysis, automatically generating high-quality financial analysis reports.

[0069] In another preferred embodiment, step S700 further includes:

[0070] Before filling the corresponding analysis content for each financial analysis item into the corresponding position of that financial analysis item, the following processing is performed for each financial analysis item: determine the target content format corresponding to the financial analysis item, the content format including at least text description, data table and chart, and convert the analysis content corresponding to the financial analysis item into the target content format.

[0071] Specifically, the content format associated with the financial analysis items can be modified. After determining the analysis content corresponding to the financial analysis items, it can be converted into the corresponding target content format according to actual needs. This can enrich the final generated financial analysis report, making the analysis results indicated in the final generated financial analysis report more intuitive and easier for users to understand the financial analysis report or conduct further analysis on the financial analysis report.

[0072] This application uses visualization technology to display the financial data corresponding to the financial analysis items in the form of charts, and supports users to conduct in-depth analysis according to different dimensions.

[0073] In a preferred embodiment, the financial analysis report corresponding to each financial analysis strategy is stored in a preset storage location in the form of a vector.

[0074] Please see Figure 2 , Figure 2 A second flowchart of an intelligent financial analysis method provided in an embodiment of this application is shown. Figure 2 As shown, the method provided in this application also includes:

[0075] S800: Receives a financial data query request input from the requesting end.

[0076] S810. Parse the financial data query request and determine the query type corresponding to the financial data query request.

[0077] The query types include financial indicator queries, financial report queries, and financial knowledge queries.

[0078] S820. If the financial data query request is a financial indicator query, then output multiple target financial analysis indicators corresponding to the financial indicator query as the response result and return them to the requesting end.

[0079] S830. If the financial data query request is a financial report query, then determine the target financial report type indicated by the financial data query request, extract the financial analysis report corresponding to the target financial report type from the preset storage location as the response result and return it to the requesting end.

[0080] S840. If the financial data query request is a financial knowledge query, then extract the query keywords corresponding to the financial data query, extract the financial knowledge corresponding to the query keywords from the financial analysis knowledge graph as the response result and return it to the requesting end.

[0081] Preferably, this application provides a human-computer dialogue function, allowing users to input corresponding financial data query requests through a human-computer dialogue interface. Upon receiving a financial data query request, the application quickly initiates a response strategy corresponding to the query type to determine the corresponding response result and feeds it back to the requesting end based on the corresponding query type. In this process, the response time of the financial data query request is greatly shortened, ensuring that the requesting end is provided with a real-time response result.

[0082] In a preferred embodiment, the method provided in this application further includes:

[0083] Acquire historical financial data generated within a preset historical time period, construct a macroeconomic forecasting model based on the historical financial data and macroeconomic indicators, determine the forecast results, and generate a decision support report based on the forecast results, financial data trends, and a preset decision-making algorithm.

[0084] Specifically, the forecast results corresponding to the macroeconomic forecasting model are the forecast values ​​corresponding to each operating indicator. This application creates a corresponding macroeconomic forecasting model through deep learning and analysis of historical financial data. By training and optimizing the macroeconomic forecasting model, the forecasting accuracy of the macroeconomic forecasting model is continuously improved, enabling the macroeconomic forecasting model to identify potential trends, patterns and correlations, thereby improving the accuracy and scientific nature of the forecasts of various operating indicators. In addition, this application also supports the generation of corresponding decision support reports.

[0085] Decision support reports can have a significant impact on corporate planning of funds, resource allocation, and strategic decisions, enabling companies to better cope with market fluctuations and competitive pressures and make more informed financial decisions.

[0086] By combining deep learning models (such as CNN, RNN, LSTM, etc.) to analyze historical data and current market conditions, we can predict the risks that enterprises may face and provide risk management advice, enabling enterprises to respond quickly and take measures to reduce the probability of risks occurring and the extent of losses.

[0087] In a preferred embodiment, the method provided in this application further includes: receiving a financial data display request sent by a requesting end, parsing the financial data display request, determining the query keywords, analysis dimensions and target display format corresponding to the target financial data, obtaining the target financial data according to the query keywords, performing dimensional analysis on the target financial data according to the analysis dimensions to obtain the target sub-financial data corresponding to each analysis dimension, converting the target sub-financial data into a target display type, and feeding back the converted target sub-financial data to the requesting end.

[0088] Specifically, the financial statement analysis system provided in this application offers the ability to intuitively display financial data and supports users in querying financial data and conducting multi-dimensional analysis.

[0089] Specifically, based on the query keywords, analysis dimensions, and target display format corresponding to the target financial data indicated in the financial data display request, the corresponding target sub-financial data under each analysis dimension is determined and displayed accordingly. This fulfills the user's need to display financial data in multiple different display formats, providing a more convenient way for users to quickly understand relevant data content. It also supports multi-dimensional analysis of financial data, assisting users in gaining a holistic and comprehensive understanding of the financial data.

[0090] Therefore, by automating the data collection, processing, and analysis process, financial analysis reports reduce human intervention during generation, enabling companies to obtain key business information more quickly and accelerate decision-making processes. At the same time, based on big data and machine learning algorithms, complex datasets can be analyzed in depth to discover hidden patterns and trends, providing management with more accurate and comprehensive business insights. Through data-driven decision-making supported by in-depth analysis, companies can reduce blind spots and subjectivity, and improve the scientific nature and accuracy of their decisions.

[0091] Based on the same application concept, this application also provides an intelligent financial analysis device corresponding to the intelligent financial analysis method provided in the above embodiments. Since the principle of the device in this application is similar to the intelligent financial analysis method in the above embodiments of this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0092] Please see Figure 3 , Figure 3 This diagram illustrates the functional modules of an intelligent financial analysis device provided in an embodiment of this application. Figure 3 As shown, the intelligent financial analysis device includes:

[0093] The data acquisition module A1 is used to collect the financial data to be analyzed corresponding to the target data source.

[0094] The analysis strategy determination module A2 is used to determine multiple financial analysis strategies based on different financial analysis needs of users;

[0095] The Financial Indicator Determination Module A3 is used to determine the financial indicators corresponding to each financial analysis strategy based on the financial data to be analyzed.

[0096] The prompt word generation module A4 is used to determine the prompt words corresponding to each financial analysis strategy based on the prompt word engineering and financial analysis template;

[0097] The model creation module A5 is used to create a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompt words, and pre-trained intelligent financial analysis model corresponding to each financial analysis strategy.

[0098] The financial analysis report generation module A6 is used to generate financial analysis reports corresponding to each financial analysis strategy by utilizing the target intelligent financial analysis model, financial analysis templates, financial indicators, and financial analysis knowledge graph.

[0099] Based on the same application concept, please refer to Figure 4 , Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Figure 4 As shown, the electronic device 900 includes a processor 910, a memory 920, and a bus 930. The memory 920 stores machine-readable instructions that can be executed by the processor 910. When the electronic device 900 is running, the processor 910 and the memory 920 communicate through the bus 930. The machine-readable instructions are executed by the processor 910 to perform the steps of any of the intelligent financial analysis methods provided in the above embodiments.

[0100] Based on the same concept, this application also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps of the intelligent financial analysis method provided in the above embodiments.

[0101] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0103] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0104] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent financial analysis method, characterized in that, The method includes: Collect the financial data to be analyzed corresponding to the target data source; Based on the different financial analysis needs of users, multiple financial analysis strategies are determined; Based on the financial data to be analyzed, determine the financial indicators corresponding to each financial analysis strategy; Based on the prompt word engineering and financial analysis template, prompt words are determined to correspond to each financial analysis strategy; Create a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompts, and pre-trained intelligent financial analysis model corresponding to each financial analysis strategy; Using the target intelligent financial analysis model, financial analysis template, financial indicators, and financial analysis knowledge graph, a financial analysis report corresponding to each financial analysis strategy is generated.

2. The method according to claim 1, characterized in that, The target data source includes multiple business system interfaces. The financial data to be analyzed is determined in the following manner: By utilizing robotic process automation (RPA) technology, corresponding financial data can be automatically identified and collected from each business system interface. The financial data is cleaned to obtain cleaned financial data. The data cleaning includes at least deduplication, error correction and missing value imputation. The cleaned financial data is validated. If the cleaned financial data passes the validation, the cleaned financial data is standardized to obtain the financial data to be analyzed. The validation includes at least consistency validation and reasonableness validation. If the cleaned financial data fails the data verification, an error data alarm is generated, which includes the error type and the location of the error data.

3. The method according to claim 1, characterized in that, The financial analysis template includes multiple financial analysis items. Specifically, prompts corresponding to each financial analysis strategy are created using the following methods: The target financial analysis item in the financial analysis template is determined, and the analysis content corresponding to the target financial analysis item is determined by the intelligent financial analysis model. For each target financial analysis item, generate corresponding content prompts using prompt word engineering; The prompts for each target financial analysis item are used to form the prompts for that financial analysis strategy.

4. The method according to claim 3, characterized in that, The target intelligent financial analysis model is created using the following method: For each financial analysis strategy, perform the following processing: From the corresponding financial indicators, determine the multiple financial analysis indicators corresponding to each target financial analysis item in the financial analysis template corresponding to the financial analysis strategy. Input multiple financial analysis indicators and content prompts corresponding to each target financial analysis item into a pre-trained intelligent financial analysis model to obtain the analysis content corresponding to each target financial analysis item. The intelligent financial analysis model is fine-tuned by using the analysis content corresponding to each target financial analysis item of each financial analysis strategy to obtain the target intelligent financial analysis model.

5. The method according to claim 1, characterized in that, The financial analysis knowledge graph was created using the following methods: Extract financial knowledge and business logic from the financial data to be analyzed, industry reports, and regulatory documents; The financial knowledge and business logic are organized in the form of a graph to form a financial knowledge graph.

6. The method according to claim 3, characterized in that, Generate financial analysis reports corresponding to each financial analysis strategy using the following methods: Determine the financial analysis indicators corresponding to each financial analysis item in the financial analysis template for this financial analysis strategy. For each financial analysis item, perform the following process: determine the method for generating the analysis content corresponding to that financial analysis item; If the analysis content corresponding to the financial analysis item is generated by the large model, then input the content prompts and financial analysis indicators corresponding to the financial analysis item into the target intelligent financial analysis model to obtain the analysis content corresponding to the financial analysis item. If the analysis content corresponding to the financial analysis item is generated by the aggregation of financial indicators, then the multiple financial analysis indicators corresponding to the financial analysis item will be directly determined as the analysis content corresponding to the financial analysis item. If the analysis content corresponding to the financial analysis item is formed by a financial analysis knowledge graph, then according to the financial analysis indicators corresponding to the financial analysis item, the corresponding financial knowledge is extracted from the financial analysis knowledge graph, and the financial knowledge is determined as the analysis content corresponding to the financial analysis item. Fill the corresponding analysis content for each financial analysis item into the corresponding position of that financial analysis item; Each completed financial analysis item forms the corresponding financial analysis report for that financial analysis strategy.

7. The method according to claim 6, characterized in that, Before filling the corresponding analysis content for each financial analysis item into the corresponding position of that financial analysis item, the method further includes: For each financial analysis item, perform the following processing: Determine the target content format for this financial analysis item. The content format should include at least text descriptions, data tables, and charts. Convert the analysis content corresponding to this financial analysis item into the target content format.

8. The method according to claim 1, characterized in that, The financial analysis report corresponding to each financial analysis strategy is stored in a preset storage location in vector form. The method further includes: Receive financial data query requests from the requesting end; The financial data query request is parsed to determine the corresponding query type, which includes financial indicator query, financial report query, and financial knowledge query. If the financial data query request is a financial indicator query, then multiple target financial analysis indicators corresponding to the financial indicator query will be output as the response result and returned to the requesting end; If the financial data query request is a financial report query, then the target financial report type indicated by the financial data query request is determined, and the financial analysis report corresponding to the target financial report type is extracted from the preset storage location as the response result and returned to the requesting end; If the financial data query request is a financial knowledge query, then the query keywords corresponding to the financial data query are extracted, and the financial knowledge corresponding to the query keywords is extracted from the financial analysis knowledge graph as the response result and returned to the requesting end.

9. The method according to claim 1, characterized in that, The method further includes: Retrieve historical financial data generated within a preset historical time period; Based on historical financial data and macroeconomic indicators, construct a macroeconomic forecasting model and determine the forecast results; A decision support report is generated based on the prediction results, the trend of financial data changes, and the preset decision-making algorithm.

10. The method according to claim 1, characterized in that, The method further includes: Receives a request from the requesting client to display financial data; Parse the financial data display request to determine the query keywords, analysis dimensions, and target display format corresponding to the target financial data; Retrieve the target financial data using the provided query keywords; The target financial data is analyzed according to the analysis dimensions to obtain the target sub-financial data corresponding to each analysis dimension. Convert the target sub-financial data into a target display type; The converted financial data for each target is then fed back to the requesting end.

11. An intelligent financial analysis device, characterized in that, The device includes: The data acquisition module collects the financial data to be analyzed corresponding to the target data source. The analysis strategy determination module is used to determine multiple financial analysis strategies based on different financial analysis needs of users; The financial indicator determination module is used to determine the financial indicators corresponding to each financial analysis strategy based on the financial data to be analyzed. The prompt word generation module is used to determine the prompt words corresponding to each financial analysis strategy based on the prompt word engineering and financial analysis template; The model creation module is used to create a target intelligent financial analysis model based on the financial analysis template, financial indicators, prompts, and pre-trained intelligent financial analysis model corresponding to each financial analysis strategy. The generation module is used to generate a financial analysis report corresponding to each financial analysis strategy by utilizing the target intelligent financial analysis model, financial analysis template, financial indicators and financial analysis knowledge graph.

12. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the machine-readable instructions are executed by the processor to perform the steps of the intelligent financial analysis method as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the intelligent financial analysis method as described in any one of claims 1 to 10.

Images