Enterprise financial statement automatic analysis system and method

Abstract

The invention discloses an enterprise financial statement automatic analysis system and method, and relates to the technical field of financial analysis, and the system comprises an enterprise and data unit, an analysis processing unit, and a result processing unit. The enterprise and data unit firstly divides and classifies the types of enterprises, extracts and collects the data of the enterprises after the enterprises are divided, and analyzes the data after the collection is completed; the analysis processing unit builds an enterprise selection model according to different enterprises through a large model technology, and builds a corresponding financial statement analysis scheme for analysis according to classified enterprise types; and the result processing unit carries out risk early warning according to the analyzed and processed result after analysis and processing. According to the invention, after analysis processing is completed, risk early warning processing is carried out on the generated result, grading is carried out during risk early warning processing, and a corresponding processing scheme is formulated, so that the accuracy during risk early warning and the efficiency and accuracy of risk processing during risk early warning are improved.

Description

Technical Field

[0001] This invention relates to the field of financial analysis technology, specifically to an automated financial statement analysis system and method for enterprises. Background Technology

[0002] Currently, corporate financial statements are prepared based on accounting standards and are written documents that systematically reflect a company's financial position on a specific date, its operating results and cash flows for a specific accounting period. They are the core basis for decision-making by stakeholders such as investors, creditors, management, and regulatory agencies. Chinese Patent No. CN112950346B discloses "An Automatic Analysis System for Corporate Financial Statements, including a conditional analysis setting module, a working status and analysis result output module, and modules for obtaining a company list, annual report, annual report data parsing, financial forms, and a company health assessment, all of which are business-related to the conditional analysis setting module and the working status and analysis result output module. It also discloses an automatic analysis method for corporate financial statements. This invention achieves automatic acquisition of company lists, automatic extraction and organization of company annual reports and financial forms, automatic acquisition of key financial item values, and automatic generation of a comprehensive company health assessment report, improving the efficiency and reducing the difficulty for investors analyzing company annual reports." Existing technologies only address the problem that current assessment processes are time-consuming and, for general investors lacking professional financial skills, can only provide limited and crude assessments of a potential company's health status, such as cash flow, profitability, growth potential, operational capabilities, and solvency. However, current financial statement analysis processes cannot pre-determine analytical models based on the company's characteristics, leading to low efficiency. Furthermore, during risk warnings, it's impossible to reasonably and accurately classify risk levels. Even if a risk is detected, accurate and appropriate measures cannot be taken to address the specific risk level. Additionally, the lack of retrospective review and verification of analysis results and processes during analysis can lead to instances of human error. Summary of the Invention

[0003] The purpose of this invention is to provide an automatic analysis system and method for enterprise financial statements to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an automatic analysis system for enterprise financial statements, comprising an enterprise and data unit, an analysis and processing unit, and a result processing unit; The enterprise and data unit first classifies the types of enterprises, then extracts and collects the enterprise data after the enterprises are classified, analyzes the data after the data collection, and then adjusts the classified enterprises. The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types. The result processing unit performs risk warnings based on the analysis and processing results, and then processes the risks after the risk warnings are completed.

[0005] Preferably, the enterprise and data unit includes an enterprise classification module, a data extraction module, and a data analysis module. The enterprise classification module classifies enterprises, and the specific classification method is as follows; (1) Classified by legal form: including individual businesses, partnerships, limited liability companies, joint-stock companies, state-owned enterprises and foreign-invested enterprises; (2) Classified by industry attributes: including agriculture, forestry, animal husbandry, fishery, mining, manufacturing, electricity, heat, gas and water production and supply, construction, wholesale and retail, transportation, warehousing and postal services, accommodation and catering, information transmission, software and information technology services, finance, real estate, business services and scientific research and technical services; (3) Classified by enterprise size: including industry, retail, software and information technology services and construction; (4) Classified by ownership: including state-owned enterprises, collective enterprises, private enterprises, foreign-invested enterprises and mixed-ownership enterprises; (5) Classified by geographical scope and financing structure: including the place of enterprise registration, business coverage and financing structure. The place of enterprise registration includes domestic enterprises / multinational enterprises, eastern enterprises / central and western enterprises, first-tier cities / third and fourth-tier cities. The business coverage includes local enterprises, regional enterprises, national enterprises and multinational enterprises. The financing structure includes equity financing, debt financing and self-raised funds.

[0006] Preferably, the data extraction module extracts data in real time through web crawlers and API interfaces, and formulates a data extraction model when extracting enterprise features. The data extraction model includes fundamental features, operational and financial features, innovation and strength features, risk and integrity features, public opinion features, correlation and network features, and strategic and behavioral features. (1) Fundamental characteristics include the number of years since establishment, registered capital, listing status, industry and region; (2) Operating and financial characteristics include revenue, profit, debt-to-equity ratio, cash flow and number of employees; (3) Innovation and strength characteristics include the number / type of patents, the number of trademarks, the proportion of R&D investment and the qualification of high-tech enterprises; (4) Risk and integrity characteristics include the number of legal proceedings, the number of times the person subject to enforcement has been dealt with, administrative penalty records, and equity freeze information; (5) Public opinion characteristics include news exposure, public opinion sentiment score, bidding activity and market share; (6) Relationship and network characteristics include shareholder chain depth, number of subsidiaries, supply chain partners and competitors; (7) Strategic and behavioral characteristics include business scope keywords, strategic cooperation events, investment and M&A events, and overseas expansion dynamics; During the process of extracting enterprise data features, the extracted feature data is cleaned, deduplicated, and formatted. Feature data from the same enterprise from different sources are associated and merged. Missing and inconsistent data are cross-validated using multi-source data. A data feature repository is built, and a regular update plan is formulated. The update plan includes a first update plan, a second update plan, and a third update plan. The first update plan is to update and adjust within one week, the second update plan is to update and adjust within half a month, and the third update plan is to update and adjust within one month.

[0007] Preferably, the data analysis module formulates a data analysis plan to analyze the extracted feature data, and when analyzing enterprises, it summarizes and analyzes the classified enterprises. The analyzed enterprise types include policy-affected enterprises, market-affected enterprises, and dual-affected enterprises. Among them, policy-affected enterprises are mainly affected by policies, market-affected enterprises are mainly affected by market competition and the macro environment, and dual-affected enterprises are mainly affected by policies and the market environment and policies. The data analysis plan includes a first analysis plan, a second analysis plan, a third analysis plan, and an anomaly data analysis plan. These plans analyze financial data characteristics, supply chain data characteristics, customer CNC data characteristics, and industrial production data characteristics. The first analysis plan analyzes the largest number of features and involves the most business departments and online experts. The second analysis plan analyzes a moderate number of features and involves a moderate number of business departments and online experts. The third analysis plan analyzes the fewest features and involves the fewest business departments and online experts. The anomaly data analysis plan analyzes anomalies generated during the analysis process separately. These anomalies include numerical anomalies, temporal anomalies, and correlation anomalies. The analysis involves collaborative processing by experts and frontline industry personnel. Problems generated during data analysis are collected, and optimization and adjustment plans are developed and implemented to optimize the data analysis plan.

[0008] Preferably, the analysis and processing unit includes a selection analysis module, an analysis monitoring module, and a periodic processing module. The selection analysis module uses large model technology to build analysis and processing schemes for enterprises affected by policies, enterprises affected by markets, and enterprises affected by both. When analyzing the financial statements of an enterprise, the key features of the enterprise are input into the large model, and the corresponding analysis and processing scheme will be automatically selected. In addition, the enterprise can also manually input features to improve the accuracy of analysis and processing. The policy-influenced enterprise analysis and processing solution references policies released by local authorities, obtains policy documents published on official websites and public accounts in real time, and conducts correlation analysis with video accounts and professors and experts who interpret policies. Among them, the market-influenced enterprise analysis and processing solution compares and analyzes industry data and macroeconomic indicators, restates financial statements using constant exchange rates, or analyzes exchange rate sensitivity, analyzes fluctuations in the macroeconomic cycle, including economic expansion and recession, analyzes and processes market supply and demand and competitive landscape, and analyzes capital market sentiment and industry technological reform and industrial upgrading. The dual-impact enterprise analysis and processing solution conducts dual analysis of policy and market environment. During the analysis process, standardized report templates are set up, and the standardized templates support the real-time generation of balance sheets, profit and loss statements and market trend tables on T+0. Vouchers are automatically generated through AI algorithms, and a comprehensive evaluation model is constructed by combining DuPont analysis and Wall score method and other multi-dimensional indicators for comprehensive evaluation. Finally, an analysis and evaluation template is generated based on the standardized template. Within the policy-influenced enterprise analysis and processing scheme and the market-influenced enterprise analysis and processing scheme, a first-level analysis scheme, a second-level analysis scheme, and a third-level analysis scheme are built. The first-level analysis scheme conducts financial statement analysis on large leading enterprises, the second-level analysis scheme conducts financial statement analysis on medium and large enterprises, and the third-level analysis scheme conducts financial statement analysis on small and medium-sized enterprises. After the analysis is completed, a data repository is built to store the results data generated during the analysis, and the stored data is labeled.

[0009] Preferably, the analysis and supervision module formulates a comprehensive supervision plan to supervise the financial statement analysis results and process. The comprehensive supervision plan includes a first supervision plan, a second supervision plan, and a third supervision plan. The first supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 12 hours after the financial statement analysis. The second supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 24 hours after the financial statement analysis. The third supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 48 hours after the financial statement analysis. A detection and analysis database is built to store the analysis results. During storage, the stored data is identified and classified. The results of the detection and analysis are analyzed separately within the first, second, and third supervision plans, including Level 1, Level 2, and Level 3. The most serious result is Level 1, the moderate result is Level 2, and the least serious result is Level 3. When the result is Level 1, the time for review and analysis is reduced by 50%; when the result is Level 2, the time is reduced by 70%; and when the result is Level 3, the time is reduced by 90%. If the review and analysis results are abnormal, the time will be reduced again. As long as the review and analysis results are still abnormal, the time will be reduced sequentially.

[0010] Preferably, the cycle processing module formulates a cycle analysis plan to conduct multi-angle investigation and analysis of the market environment and policies within a specified cycle, determines the core indicators that need to be monitored regularly based on the industry in which the enterprise is located, formulates dynamic analysis methods, and regularly uses the PEST framework to organize and analyze the latest data and events collected, and assess their specific impact on the enterprise. At the same time, the size of enterprises is classified. When the enterprise size is classified as large, the core indicators of the industry can be tracked in all aspects. When the enterprise size is classified as small to medium, three to five core indicators related to the industry can be tracked. In addition, qualitative judgments and expectations of entrepreneurs and consumers about the future are collected. After the periodic survey and analysis is completed, the enterprise size, enterprise type and enterprise characteristics are readjusted.

[0011] Preferably, the result processing unit includes a risk warning module and a risk processing module. The risk warning module formulates a general risk warning plan to conduct risk warning processing for enterprises. The general risk warning plan includes a first risk warning plan, a second risk warning plan, and a third risk warning plan. The first risk warning plan has the highest risk warning level, the second risk warning plan has a medium risk warning level, and the third risk warning plan has the lowest risk warning level. When classifying risk levels, the results of published policy documents and market surveys are taken into account.

[0012] The risk handling module formulates a risk handling plan when a risk is generated. The risk handling plan includes a first handling plan, a second handling plan, and a third handling plan. The first handling plan handles the problems generated by the first risk warning plan, the second handling plan handles the problems generated by the second risk warning plan, and the third handling plan handles the problems generated by the third handling plan. The first handling plan is implemented and adjusted immediately, the second handling plan can be implemented and adjusted later, and the third handling plan can be implemented and adjusted later.

[0013] An automated method for analyzing corporate financial statements includes the following steps; Step 1: The enterprises and data units first classify the types of enterprises. After the enterprises are classified, the enterprise data is extracted and collected. After the data is collected, the data is analyzed, and then the classified enterprises are adjusted. Step 2: The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types; Step 3: After analysis and processing, the result processing unit performs risk warning based on the results of the analysis and processing, and then processes the risk after the risk warning is completed.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention classifies enterprises and formulates corresponding financial statement analysis schemes based on the characteristics of the classified enterprises. Through customized processing, it improves the efficiency and accuracy of analysis and processing. During the analysis process, the analysis results are reviewed, tested, and checked to ensure the accuracy of the financial statements and avoid cheating during the analysis process, which could lead to an inaccurate final structure. 2. After the analysis and processing are completed, the present invention performs risk warning processing on the generated results, classifies the risk warnings, and formulates corresponding processing plans to improve the accuracy of risk warnings and the efficiency and accuracy of risk handling during risk warnings. Attached Figure Description

[0015] Figure 1 A system block diagram is provided for embodiments of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1 Please see Figure 1 The present invention provides a technical solution: an automatic analysis system for enterprise financial statements, comprising an enterprise and data unit, an analysis and processing unit, and a result processing unit; The enterprise and data unit first classifies the types of enterprises, then extracts and collects the enterprise data after the enterprises are classified, analyzes the data after the data collection, and then adjusts the classified enterprises. The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types. The result processing unit performs risk warnings based on the analysis and processing results, and then processes the risks after the risk warnings are completed.

[0018] The enterprise and data unit includes an enterprise classification module, a data extraction module, and a data analysis module. The enterprise classification module classifies enterprises, and the specific classification method is as follows; (1) Classified by legal form: including individual businesses, partnerships, limited liability companies, joint-stock companies, state-owned enterprises and foreign-invested enterprises; (2) Classified by industry attributes: including agriculture, forestry, animal husbandry, fishery, mining, manufacturing, electricity, heat, gas and water production and supply, construction, wholesale and retail, transportation, warehousing and postal services, accommodation and catering, information transmission, software and information technology services, finance, real estate, business services and scientific research and technical services; (3) Classified by enterprise size: including industry, retail, software and information technology services and construction; (4) Classified by ownership: including state-owned enterprises, collective enterprises, private enterprises, foreign-invested enterprises and mixed-ownership enterprises; (5) Classified by geographical scope and financing structure: including the place of enterprise registration, business coverage and financing structure. The place of enterprise registration includes domestic enterprises / multinational enterprises, eastern enterprises / central and western enterprises, first-tier cities / third and fourth-tier cities. The business coverage includes local enterprises, regional enterprises, national enterprises and multinational enterprises. The financing structure includes equity financing (relying on shareholder investment and IPO fundraising), debt financing (relying on bank loans and bond issuance) and self-raised funds (relying on own funds and operating cash flow).

[0019] The data extraction module extracts data in real time through web crawlers and API interfaces. When extracting enterprise features, a data extraction model is formulated, which includes fundamental features, operational and financial features, innovation and strength features, risk and integrity features, public opinion features, correlation and network features, and strategic and behavioral features. (1) Fundamental characteristics include the number of years since establishment, registered capital, listing status, industry and region; (2) Operating and financial characteristics include revenue, profit, debt-to-equity ratio, cash flow and number of employees; (3) Innovation and strength characteristics include the number / type of patents, the number of trademarks, the proportion of R&D investment and the qualification of high-tech enterprises; (4) Risk and integrity characteristics include the number of legal proceedings, the number of times the person subject to enforcement has been dealt with, administrative penalty records, and equity freeze information; (5) Public opinion characteristics include news exposure, public opinion sentiment score, bidding activity and market share; (6) Relationship and network characteristics include shareholder chain depth, number of subsidiaries, supply chain partners and competitors; (7) Strategic and behavioral characteristics include business scope keywords, strategic cooperation events, investment and M&A events, and overseas expansion dynamics; During the process of extracting enterprise data features, the extracted feature data is cleaned, deduplicated, and formatted. Feature data from the same enterprise from different sources are associated and merged. Missing and inconsistent data are cross-validated using multi-source data. A data feature repository is built, and a regular update plan is formulated. The update plan includes a first update plan, a second update plan, and a third update plan. The first update plan is to update and adjust within one week, the second update plan is to update and adjust within half a month, and the third update plan is to update and adjust within one month.

[0020] The data analysis module formulates a data analysis plan to analyze the extracted feature data. When analyzing enterprises, it summarizes and analyzes the categorized enterprises. The analyzed enterprise types include policy-affected enterprises, market-affected enterprises, and dual-affected enterprises. Policy-affected enterprises are mainly affected by policies, market-affected enterprises are mainly affected by market competition and the macro environment, and dual-affected enterprises are mainly affected by policies and the market environment and policies. The data analysis plan includes a first analysis plan, a second analysis plan, a third analysis plan, and an anomaly data analysis plan. These plans analyze financial data characteristics, supply chain data characteristics, customer CNC data characteristics, and industrial production data characteristics. The first analysis plan analyzes the most features and involves the largest number of business departments and online experts. The second analysis plan analyzes a moderate number of features and involves a moderate number of business departments and online experts. The third analysis plan analyzes the fewest features and involves the fewest business departments and online experts. The anomaly data analysis plan analyzes anomalies generated during the analysis process separately. These anomalies include numerical anomalies (e.g., unusually high reimbursement amounts), temporal anomalies (e.g., sudden drops in sales), and correlational anomalies (e.g., low-credit customers placing large orders). The analysis involves collaborative processing by experts and frontline industry personnel. Problems generated during data analysis are collected, and optimization and adjustment plans are developed and implemented to optimize the data analysis plan.

[0021] The analysis and processing unit includes a selection analysis module, an analysis monitoring module, and a periodic processing module. The selection analysis module uses large model technology to build analysis and processing schemes for enterprises affected by policies, enterprises affected by markets, and enterprises affected by both. When analyzing a company's financial statements, the key features of the company are input into the large model, and the corresponding analysis and processing scheme will be automatically selected. In addition, the company can also manually input features to improve the accuracy of the analysis and processing. The policy-influenced enterprise analysis and processing solution references policies released by local authorities, obtains policy documents published on official websites and public accounts in real time, and conducts correlation analysis with video accounts and professors and experts who interpret policies. Among them, the market-influenced enterprise analysis and processing solution compares and analyzes industry data and macroeconomic indicators (such as GDP growth rate and PMI), restates reports using constant exchange rates, or analyzes exchange rate sensitivity, analyzes fluctuations in the macroeconomic cycle, including economic expansion and recession, analyzes and processes market supply and demand and competitive landscape, and analyzes capital market sentiment and industry technological reform and industrial upgrading. The dual-impact enterprise analysis and processing solution conducts dual analysis of policy and market environment. During the analysis process, standardized report templates are set up, and the standardized templates support the real-time generation of balance sheets, profit and loss statements and market trend tables on T+0. Vouchers are automatically generated through AI algorithms, and a comprehensive evaluation model is constructed by combining DuPont analysis and Wall score method and other multi-dimensional indicators for comprehensive evaluation. Finally, an analysis and evaluation template is generated based on the standardized template. Within the policy-influenced enterprise analysis and processing scheme and the market-influenced enterprise analysis and processing scheme, a first-level analysis scheme, a second-level analysis scheme, and a third-level analysis scheme are built. The first-level analysis scheme conducts financial statement analysis on large leading enterprises, the second-level analysis scheme conducts financial statement analysis on medium and large enterprises, and the third-level analysis scheme conducts financial statement analysis on small and medium-sized enterprises. After the analysis is completed, a data repository is built to store the results data generated during the analysis, and the stored data is labeled.

[0022] The analysis and supervision module formulates a comprehensive supervision plan to supervise the results and process of financial statement analysis. The comprehensive supervision plan includes a first supervision plan, a second supervision plan, and a third supervision plan. The first supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 12 hours after the financial statement analysis. The second supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 24 hours after the financial statement analysis. The third supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 48 hours after the financial statement analysis. A detection and analysis database is built to store the results of the analysis. During storage, the stored data is identified and classified. The results of the detection and analysis are analyzed separately within the first, second, and third supervision plans, including Level 1, Level 2, and Level 3. The most serious result is Level 1, the moderate result is Level 2, and the least serious result is Level 3. When the result is Level 1, the time for review and analysis is reduced by 50%; when the result is Level 2, the time is reduced by 70%; and when the result is Level 3, the time is reduced by 90%. If the review and analysis results are abnormal, the time will be reduced again. As long as the review and analysis results are still abnormal, the time will be reduced sequentially.

[0023] The cycle processing module formulates a cycle analysis plan. Within the specified cycle, it conducts multi-angle investigation and analysis of the market environment and policies. Based on the industry in which the enterprise operates, it determines the core indicators that need to be monitored regularly, formulates dynamic analysis methods, and regularly (such as monthly, quarterly, and annually) uses the PEST framework to organize and analyze the latest data and events collected, and assess their specific impact on the enterprise. At the same time, the size of enterprises is classified. When the enterprise size is classified as large, the core indicators of the industry can be tracked in all aspects. When the enterprise size is classified as small to medium, three to five core indicators related to the industry can be tracked. In addition, qualitative judgments and expectations of entrepreneurs and consumers about the future are collected. After the periodic survey and analysis is completed, the enterprise size, enterprise type and enterprise characteristics are readjusted.

[0024] The result processing unit includes a risk warning module and a risk processing module. The risk warning module formulates a general risk warning plan to conduct risk warning processing for enterprises. The general risk warning plan includes a first risk warning plan, a second risk warning plan, and a third risk warning plan. The first risk warning plan has the highest risk warning level, the second risk warning plan has a medium risk warning level, and the third risk warning plan has the lowest risk warning level. When classifying risk levels, the results of published policy documents and market surveys are taken into account.

[0025] The risk handling module formulates a risk handling plan when a risk is generated. The risk handling plan includes a first handling plan, a second handling plan, and a third handling plan. The first handling plan handles the problems generated by the first risk warning plan, the second handling plan handles the problems generated by the second risk warning plan, and the third handling plan handles the problems generated by the third handling plan. The first handling plan is implemented and adjusted immediately, the second handling plan can be implemented and adjusted later, and the third handling plan can be implemented and adjusted later. During the process, we comprehensively identify potential financial risks faced by the company through financial statement analysis, industry research and management discussions, quantify the size and probability of the risks, establish a regular risk reporting system, and report the risk status and the effectiveness of the countermeasures to the management and the board of directors. (1) Simultaneously conduct cash budget management, prepare detailed monthly, weekly and even daily cash budgets, strictly monitor the inflow and outflow of funds, establish emergency funds to retain a certain proportion of cash or hold easily convertible short-term assets (such as money market funds, treasury bonds, etc.), and diversify financing channels by developing multiple channels such as bill financing, financial leasing, supply chain finance and equity financing. (2) Optimize the capital structure to determine a reasonable debt-to-equity ratio, maintain a balance between equity capital and debt capital, match debt maturity, and ensure that long-term assets (such as fixed assets) are supported by long-term liabilities (such as long-term loans and bonds), and short-term assets are supported by short-term liabilities. Avoid short-term loans for long-term investments. Before the debt matures, plan new financing sources in advance to ensure the continuity of funds. In extreme cases, negotiate with creditors to convert debt into equity in order to reduce the debt ratio. (3) Implement cost control and lean management, reduce fixed and variable costs, adjust pricing strategies, flexibly adjust pricing strategies according to market conditions and cost structure, analyze cost, volume and profit, conduct break-even point analysis regularly, and clarify the break-even point and safety margin. (4) Establish and improve the internal control system, improve the approval process, separation of duties and internal audit, etc., which can effectively prevent fraud and operational errors, include all business activities of the enterprise in the budget, and manage risks through pre-planning, in-process control and post-analysis.

[0026] Example 2 An automated method for analyzing corporate financial statements includes the following steps; Step 1: The enterprises and data units first classify the types of enterprises. After the enterprises are classified, the enterprise data is extracted and collected. After the data is collected, the data is analyzed, and then the classified enterprises are adjusted. Step 2: The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types; Step 3: After analysis and processing, the result processing unit performs risk warning based on the results of the analysis and processing, and then processes the risk after the risk warning is completed.

[0027] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0028] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automated financial statement analysis system for enterprises, characterized in that... It includes enterprise and data units, analysis and processing units, and results processing units; The enterprise and data unit first classifies the types of enterprises, then extracts and collects the enterprise data after the enterprises are classified, analyzes the data after the data collection, and then adjusts the classified enterprises. The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types. The result processing unit performs risk warnings based on the analysis and processing results, and then processes the risks after the risk warnings are completed.

2. The automatic analysis system for enterprise financial statements according to claim 1, characterized in that: The enterprise and data unit includes an enterprise classification module, a data extraction module, and a data analysis module. The enterprise classification module classifies enterprises, and the specific classification method is as follows; (1) Classified by legal form: including individual businesses, partnerships, limited liability companies, joint-stock companies, state-owned enterprises and foreign-invested enterprises; (2) Classified by industry attributes: including agriculture, forestry, animal husbandry, fishery, mining, manufacturing, electricity, heat, gas and water production and supply, construction, wholesale and retail, transportation, warehousing and postal services, accommodation and catering, information transmission, software and information technology services, finance, real estate, business services and scientific research and technical services; (3) Classified by enterprise size: including industry, retail, software and information technology services and construction; (4) Classified by ownership: including state-owned enterprises, collective enterprises, private enterprises, foreign-invested enterprises and mixed-ownership enterprises; (5) Classified by geographical scope and financing structure: including the place of enterprise registration, business coverage and financing structure. The place of enterprise registration includes domestic enterprises / multinational enterprises, eastern enterprises / central and western enterprises, first-tier cities / third and fourth-tier cities. The business coverage includes local enterprises, regional enterprises, national enterprises and multinational enterprises. The financing structure includes equity financing, debt financing and self-raised funds.

3. The automatic analysis system for enterprise financial statements according to claim 2, characterized in that: The data extraction module extracts data in real time through web crawlers and API interfaces. When extracting enterprise features, a data extraction model is formulated, which includes fundamental features, operational and financial features, innovation and strength features, risk and integrity features, public opinion features, correlation and network features, and strategic and behavioral features. (1) Fundamental characteristics include the number of years since establishment, registered capital, listing status, industry and region; (2) Operating and financial characteristics include revenue, profit, debt-to-equity ratio, cash flow and number of employees; (3) Innovation and strength characteristics include the number / type of patents, the number of trademarks, the proportion of R&D investment and the qualification of high-tech enterprises; (4) Risk and integrity characteristics include the number of legal proceedings, the number of times the person subject to enforcement has been dealt with, administrative penalty records, and equity freeze information; (5) Public opinion characteristics include news exposure, public opinion sentiment score, bidding activity and market share; (6) Relationship and network characteristics include shareholder chain depth, number of subsidiaries, supply chain partners and competitors; (7) Strategic and behavioral characteristics include business scope keywords, strategic cooperation events, investment and M&A events, and overseas expansion dynamics; During the process of extracting enterprise data features, the extracted feature data is cleaned, deduplicated, and formatted. Feature data from the same enterprise from different sources are associated and merged. Missing and inconsistent data are cross-validated using multi-source data. A data feature repository is built, and a regular update plan is formulated. The update plan includes a first update plan, a second update plan, and a third update plan. The first update plan is to update and adjust within one week, the second update plan is to update and adjust within half a month, and the third update plan is to update and adjust within one month.

4. The automatic analysis system for enterprise financial statements according to claim 3, characterized in that: The data analysis module formulates a data analysis plan to analyze the extracted feature data. When analyzing enterprises, it summarizes and analyzes the categorized enterprises. The analyzed enterprise types include policy-affected enterprises, market-affected enterprises, and dual-affected enterprises. Policy-affected enterprises are mainly affected by policies, market-affected enterprises are mainly affected by market competition and the macro environment, and dual-affected enterprises are mainly affected by policies and the market environment and policies. The data analysis plan includes a first analysis plan, a second analysis plan, a third analysis plan, and an anomaly data analysis plan. These plans analyze financial data characteristics, supply chain data characteristics, customer CNC data characteristics, and industrial production data characteristics. The first analysis plan analyzes the largest number of features and involves the most business departments and online experts. The second analysis plan analyzes a moderate number of features and involves a moderate number of business departments and online experts. The third analysis plan analyzes the fewest features and involves the fewest business departments and online experts. The anomaly data analysis plan analyzes anomalies generated during the analysis process separately. These anomalies include numerical anomalies, temporal anomalies, and correlation anomalies. The analysis involves collaborative processing by experts and frontline industry personnel. Problems generated during data analysis are collected, and optimization and adjustment plans are developed and implemented to optimize the data analysis plan.

5. The automatic analysis system for enterprise financial statements according to claim 4, characterized in that: The analysis and processing unit includes a selection analysis module, an analysis monitoring module, and a periodic processing module. The selection analysis module uses large model technology to build analysis and processing schemes for enterprises affected by policies, enterprises affected by markets, and enterprises affected by both. When analyzing a company's financial statements, the key features of the company are input into the large model, and the corresponding analysis and processing scheme will be automatically selected. In addition, the company can also manually input features to improve the accuracy of the analysis and processing. The policy-influenced enterprise analysis and processing solution references policies released by local authorities, obtains policy documents published on official websites and public accounts in real time, and conducts correlation analysis with video accounts and professors and experts who interpret policies. Among them, the market-influenced enterprise analysis and processing solution compares and analyzes industry data and macroeconomic indicators, restates financial statements using constant exchange rates, or analyzes exchange rate sensitivity, analyzes fluctuations in the macroeconomic cycle, including economic expansion and recession, analyzes and processes market supply and demand and competitive landscape, and analyzes capital market sentiment and industry technological reform and industrial upgrading. The dual-impact enterprise analysis and processing solution conducts dual analysis of policy and market environment. During the analysis process, standardized report templates are set up, and the standardized templates support the real-time generation of balance sheets, profit and loss statements and market trend tables on T+0. Vouchers are automatically generated through AI algorithms, and a comprehensive evaluation model is constructed by combining DuPont analysis and Wall score method and other multi-dimensional indicators for comprehensive evaluation. Finally, an analysis and evaluation template is generated based on the standardized template. Within the policy-influenced enterprise analysis and processing scheme and the market-influenced enterprise analysis and processing scheme, a first-level analysis scheme, a second-level analysis scheme, and a third-level analysis scheme are built. The first-level analysis scheme conducts financial statement analysis on large leading enterprises, the second-level analysis scheme conducts financial statement analysis on medium and large enterprises, and the third-level analysis scheme conducts financial statement analysis on small and medium-sized enterprises. After the analysis is completed, a data repository is built to store the results data generated during the analysis, and the stored data is labeled.

6. The automatic analysis system for enterprise financial statements according to claim 5, characterized in that: The analysis and supervision module formulates a comprehensive supervision plan to supervise the results and process of financial statement analysis. The comprehensive supervision plan includes a first supervision plan, a second supervision plan, and a third supervision plan. The first supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 12 hours after the financial statement analysis. The second supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 24 hours after the financial statement analysis. The third supervision plan involves experts conducting a review and analysis of the analyzed financial statements within 48 hours after the financial statement analysis. A detection and analysis database is built to store the results of the analysis. During storage, the stored data is identified and classified. The results of the detection and analysis are analyzed separately within the first, second, and third supervision plans, including Level 1, Level 2, and Level 3. The most serious result is Level 1, the moderate result is Level 2, and the least serious result is Level 3. When the result is Level 1, the time for review and analysis is reduced by 50%; when the result is Level 2, the time is reduced by 70%; and when the result is Level 3, the time is reduced by 90%. If the review and analysis results are abnormal, the time will be reduced again. As long as the review and analysis results are still abnormal, the time will be reduced sequentially.

7. The automatic analysis system for enterprise financial statements according to claim 6, characterized in that: The cycle processing module formulates a cycle analysis plan. Within a specified cycle, it conducts multi-faceted investigations and analyses of the market environment and policies. Based on the industry in which the enterprise operates, it determines the core indicators that need to be monitored regularly, formulates dynamic analysis methods, and regularly uses the PEST framework to organize and analyze the latest data and events collected, and assesses their specific impact on the enterprise. At the same time, the size of enterprises is classified. When the enterprise size is classified as large, the core indicators of the industry can be tracked in all aspects. When the enterprise size is classified as small to medium, three to five core indicators related to the industry can be tracked. In addition, qualitative judgments and expectations of entrepreneurs and consumers about the future are collected. After the periodic survey and analysis is completed, the enterprise size, enterprise type and enterprise characteristics are readjusted.

8. The automatic analysis system for enterprise financial statements according to claim 1, characterized in that: The result processing unit includes a risk warning module and a risk processing module. The risk warning module formulates a general risk warning plan to conduct risk warning processing for enterprises. The general risk warning plan includes a first risk warning plan, a second risk warning plan, and a third risk warning plan. The first risk warning plan has the highest risk warning level, the second risk warning plan has a medium risk warning level, and the third risk warning plan has the lowest risk warning level. When classifying risk levels, the results of published policy documents and market surveys are taken into account.

9. The risk handling module formulates a risk handling plan when a risk is generated, and the risk handling plan includes a first handling plan, a second handling plan, and a third handling plan. The first handling plan handles the problems generated by the first risk warning plan, the second handling plan handles the problems generated by the second risk warning plan, and the third handling plan handles the problems generated by the third handling plan. The first handling plan is implemented and adjusted immediately, the second handling plan can be implemented and adjusted later, and the third handling plan can be implemented and adjusted later.

10. A method for automatically analyzing enterprise financial statements, characterized in that: The automatic analysis method for enterprise financial statements described in claim 8 includes the following steps: Step 1: The enterprises and data units first classify the types of enterprises. After the enterprises are classified, the enterprise data is extracted and collected. After the data is collected, the data is analyzed, and then the classified enterprises are adjusted. Step 2: The analysis and processing unit uses large model technology to build enterprise selection models based on different enterprises, and builds corresponding financial statement analysis schemes for analysis based on the classified enterprise types; Step 3: After analysis and processing, the result processing unit performs risk warning based on the results of the analysis and processing, and then processes the risk after the risk warning is completed.

Images