Intelligent financial statement automatic generation platform

Abstract

The invention discloses an intelligent financial statement automatic generation platform, and relates to the technical field of financial statement generation, and the platform comprises a data integration module which is used for connecting and synchronizing a data source, the data source comprises a bank transaction record, an enterprise resource planning system and an electronic invoice system, and data is grabbed through connection configuration of a data interface. According to the intelligent financial statement automatic generation platform, classification rules of financial subjects are customized by a user, so that the platform can adapt to specific financial statement requirements and business characteristics of different enterprises, the flexibility and applicability of the system are improved, statement customization is supported through the platform, financial statements of different versions are provided, and the user experience is improved. According to the method, different requirements of different users on financial information are met, the flexibility and applicability of the system are improved, a temporary version exclusive to the current requirement is formulated based on a customizable mechanism of a classification rule, and more targeted financial support is provided for decision and management of enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of financial report generation, and in particular to an intelligent financial report automatic generation platform. Background Art

[0002] With the continuous development of corporate business and the acceleration of digitalization, the generation and management of financial statements are crucial to corporate decision-making and operations; the traditional method of generating financial statements mainly relies on manual operations, which has many problems and is difficult to meet the needs of modern enterprises; on the one hand, the traditional financial statement generation process usually requires financial personnel to manually collect data from multiple data sources. This process is not only time-consuming and labor-intensive, but also prone to data omissions, errors or inconsistencies; on the other hand, in the classification of financial accounts and report generation, traditional methods mainly rely on manual judgment and experience, lacking scientificity and accuracy; at the same time, traditional report generation methods often lack flexibility and customization capabilities, making it difficult to meet the diverse needs of different users for financial statements. Summary of the invention

[0003] To solve the above technical problems, the present invention is implemented through the following technical solutions: an intelligent financial statement automatic generation platform, including a data integration module, which is used to connect and synchronize data sources, including bank transaction records, enterprise resource planning systems and electronic invoice systems, and to capture data by configuring the connection of data interfaces. The module supports data integration methods including API, direct database connection and file import; The identification and classification module is connected to the data integration module and is used to analyze and identify the correlation between the intermediate data and the financial subjects through machine learning algorithms, obtain the financial related subjects, and archive the subject attribution according to the preset financial subjects; The report generation module is connected with the identification and classification module and is used to generate financial reports by summarizing the classified accounts according to the enterprise financial reporting standards and the user-defined report format. It also supports report customization and provides different versions of financial reports. The regulatory verification module is connected to the report generation module and performs regulatory compliance checks on financial statements through the built-in financial regulatory library to verify whether the generated financial statements comply with the preset financial regulations and financial standards.

[0004] Preferably, the data integration module includes: The interface configuration unit is used to set and manage the connection configuration of different data sources. The connection configuration includes authentication information, data capture frequency and data type. It also supports custom configuration to adapt to different data sources and business needs. A synchronization control unit is connected to the interface configuration unit and is used to control the synchronization process of the data source to ensure timely data update and synchronization accuracy, obtain log records by real-time monitoring of the state and performance of data source synchronization, and handle exceptions that occur during the synchronization process according to the log records; The data cache unit is connected to the synchronization control unit and is used to temporarily store the data synchronized from the data source, recorded as the original data, and manage the temporary storage and access of the data through an efficient cache strategy, thereby reducing the waiting time of the data processing module and improving the overall performance of the system; The data processing unit is connected to the data cache unit, receives the raw data generated by the data cache unit, and cleans the raw data, including removing duplicate records, format conversion, missing value processing and outlier detection, obtains intermediate data in a unified format, and sends the intermediate data to the recognition and classification module; The interface configuration unit is used to set and manage the connection configuration of different data sources as follows: For data sources that require authentication, collect the corresponding authentication information; determine the data capture frequency based on business needs and the update frequency of the data source; analyze the format and structure of the data in the data source, determine the data type, and establish a connection with the data source based on the authentication information, data capture frequency and data type; The process of custom configuration is as follows: Identify the business needs of the data source based on the connected data source, analyze the data source with configuration requirements, obtain the characteristics of the data source, including data structure, update mechanism and authentication method, and determine how to effectively connect to and obtain data from the data source; design a custom configuration plan based on programming tools and programming techniques according to business needs and data source characteristics; the custom configuration plan includes authentication method, data capture frequency, data type processing, and data filtering and screening; The process of getting the log record: Clarify the synchronization information of the data source that needs to be recorded, including data source connection information, synchronization start time and end time, data capture quantity and progress, performance indicators during synchronization, error information and warning information; adopt the file log recording method, insert the log recording logic in the code of the synchronization control unit, record the end time and performance indicators at the end of synchronization, and record the error information or warning information when an error or warning occurs; and configure the log level based on the synchronization information of the data source that needs to be recorded, so as to complete the log recording process; The process of handling exceptions that occur during synchronization according to log records: Use scripts to scan log records regularly to find abnormal information, generate alarms based on the abnormal information, obtain error information in log records based on the alarms and perform abnormal analysis to obtain the cause of the abnormality, take processing measures based on the cause of the abnormality, obtain the exception processing results based on the processing measures of the cause of the abnormality, and record the exception processing results in the log. The exception processing results include the measures taken, processing time and result status, and complete the processing of exceptions that occur during the synchronization process according to the log records, which helps with subsequent problem troubleshooting and system optimization.

[0005] Preferably, the identification and classification module includes: A feature extraction unit is used to identify and extract key financial features from the preprocessed intermediate data. The key financial features include amount, date, and event description, and use a machine learning algorithm to analyze the correlation between the key financial features and financial accounts to identify the financial related accounts corresponding to the data items; The classification decision unit is connected to the feature extraction unit and is used to classify the intermediate data into corresponding financial accounts according to the extracted key financial features and the obtained financial related accounts, and generate account attribution; it is also used to judge the account attribution of each record through the pre-trained classification model, obtain the judgment result, and support users to customize the classification rules of financial accounts to meet different financial reporting requirements; The data archiving unit is connected to the classification decision unit and is used to archive and store the classified accounts in the financial database by account, obtain financial structure data, ensure the orderliness and retrievability of the data, support efficient data query and update operations, and provide an accurate data source for report generation.

[0006] Preferably, the pre-trained classification model determines the subject attribution of each record to obtain a determination result: The classification model is deployed to the classification decision unit of the recognition and classification module, and the prepared input data is input into the classification model for classification prediction, and the classification result is output. The classification result represents the probability that each record belongs to a different financial account, and the probability of each financial account is compared with the corresponding determined probability threshold to determine the account attribution judgment result of each record, wherein during the comparison: if the probability of the financial account is greater than or equal to the probability threshold, the record is classified into this financial account; if the probability of the financial account is less than the probability threshold, the record is marked for manual review; The process of building a classification model is: Clarify the problem of financial account classification, obtain account categories, clarify the characteristic indicators of classification, which include classification accuracy and recall rate, and use characteristic indicators as the optimization target of the model. Obtain financial data records based on the enterprise financial system and database as well as the public financial statement data set. The financial data records contain different types of financial transaction information and have corresponding financial account annotations; clean the financial data records, remove outliers, erroneous data and duplicate data, and obtain a financial data set. The financial data set is divided into a training set, a validation set and a test set. According to the account category of financial account classification, a deep neural network is selected as the architecture of the classification model. The selected deep neural network is trained using the training set data. The trained model is evaluated using the validation set data to obtain the evaluation results. The parameters and hyperparameters of the model are adjusted according to the evaluation results to obtain the initial model. The initial model is tested using the test set data to verify the performance of the model on unseen data, so that the verification results of the model meet the optimization goals, and finally the classification model is obtained.

[0007] Preferably, the process of the user defining the classification rules of financial accounts: First, clarify the financial reporting requirements, determine which specific financial account classifications are required and the classification standards; according to the financial reporting requirements, design customized financial account classification rules and define them based on financial characteristics: Transaction amount range: records are classified into different accounts according to the size of the transaction amount; Transaction object type: classified according to the nature of the transaction object; Business type: classified according to the specific type of business; Convert the designed classification rules into executable configuration files, and input the configuration files through the rule configuration interface provided by the system; test the custom classification rules with known financial data before application; check whether the rules can correctly classify the records into the expected subjects; if not, optimize and adjust the classification rules to obtain clear custom classification rules; then apply the tested classification rules to financial data processing; and in the subsequent use process, regularly check the validity of the classification rules, optimize and adjust them as the business changes and financial reporting requirements, and update and maintain the custom classification rules in a timely manner.

[0008] Preferably, the process of regularly checking the effectiveness of classification rules and optimizing and adjusting them as business changes and financial reporting requirements change is as follows: Determine the inspection cycle of classification rules based on the business characteristics of the enterprise and the importance of financial reports. For example, for enterprises with rapidly changing businesses, inspections can be conducted once a month, and for enterprises with relatively stable businesses, inspections can be conducted once a quarter or half a year. Collect feedback information on the effectiveness of classification rules from three channels: financial personnel, management and auditors; inspect and analyze recent financial data. The inspection and analysis includes inspecting the distribution of data, comparing data from different periods and analyzing abnormal data to obtain financial data analysis results; evaluate the effectiveness of classification rules based on the collected feedback information and financial data analysis results to obtain evaluation results. The evaluation content includes accuracy, consistency and adaptability. If problems are found in the classification rules, optimize and adjust them based on the evaluation results. Optimization and adjustment include modifying the rules, retraining the model, and testing and verification. For modifying the rules: optimize the classification rules based on the feedback information and analysis results. Make specific modifications, for example, adjust the classification standards, add new classification items, and modify the scope of application of the rules; for retraining models: new data can be used to retrain the model to improve the accuracy and adaptability of classification; for testing and verification: before implementing new classification rules, conduct sufficient testing and verification. A portion of the data can be used for simulated classification to check the effectiveness of the new rules. If necessary, relevant personnel can be invited to try them out and collect their feedback. Subsequently, the optimized classification rules are determined, and training materials and guidelines are generated based on the optimized classification rules. Communicate and train financial personnel and management, and after the classification rules are adjusted, continuously monitor the optimized classification rules and their effects, regularly collect feedback information, analyze financial data, continue to evaluate the effectiveness of the classification rules, and make timely adjustments and optimizations based on business changes and new needs to obtain new classification rules to ensure that the classification rules can always meet the company's financial reporting needs.

[0009] Preferably, the report generation module includes: The summary calculation unit is used to calculate the summary amount of each account based on the financial structure data, and generate a report template in combination with the enterprise financial report standard. The report template includes profit summary, asset-liability matching and cash liquidity analysis; The report construction unit is connected to the summary calculation unit and is used to construct financial reports based on the report template and report format generated by the summary calculation unit. It supports multiple report formats and customized content, chart insertion, annotation addition and ratio calculation to improve the readability and information value of the report.

[0010] Preferably, the process of constructing financial statements by the report construction unit is as follows: Parse the report template, clarify the report structure, report format and data fields required for the report, identify the formulas and calculation logic in the report, extract the corresponding financial account data from the financial structure data according to the requirements of the report template, obtain the extracted information, calculate the data fields that need to be calculated according to the formulas and calculation logic in the report template, obtain the calculated data, and fill the extracted information and calculated data into the corresponding positions of the report template. Format and layout the title, subtitle and comments in the report to make them clear and easy to read. Adjust and beautify the format of the report according to the user's requirements and aesthetic standards, and use charts and graphics to adjust the report layout to ensure that the data display is reasonable and clear, and enhance the report The financial statements are obtained by checking the readability and intuitiveness of the financial statements. The constructed financial statements are then manually reviewed and verified to check the accuracy of the data, the correctness of the formulas and the completeness of the statements. If any errors or problems are found, they are corrected and adjusted in a timely manner to ensure the quality and reliability of the financial statements. After the financial statements are reviewed and verified, they are output to the format required by the user, and report customization is supported to provide different versions of financial statements according to user needs. Different versions of financial statements include detailed version, summary version and specific department version to meet the needs of different users. Through the above process, the report construction unit can construct accurate, beautiful and professional financial statements based on the report templates and financial structure data generated by the summary calculation unit.

[0011] Preferably, the process of providing different versions of financial statements based on report customization according to user needs is as follows: Detailed financial statement generation process: Obtain financial structure data, arrange the financial structure data according to the requirements of financial statements, organize and classify the financial structure data, fill the organized financial structure data one by one into the corresponding positions according to the template of the detailed report, ensure that each financial account has a detailed display, carry out calculations for the fields that need to be calculated, obtain the calculation results, and display the calculation process and results in the report in the form of annotations. In this way, some complex financial accounts or transactions can be explained so that users can better understand the report content, set the font, font size and color to make the report easy to read, and use different colors to distinguish different financial accounts and important information, and add bar charts and line charts to show the trends and changes of financial data, adjust the page layout, ensure the overall appearance and professionalism of the report, and invite auditors to review and verify the detailed financial statements. The review and verification includes checking the accuracy of the data, the correctness of the calculation and the standardization of the format to ensure that the report meets the financial specifications and requirements. Then, the approved detailed financial statements are output based on the format required by the user, and the reports are sent to the users via email or file sharing platforms to complete the delivery. Summary financial statement generation process: Obtain the key financial features generated in the feature extraction unit, summarize and calculate the key financial features, obtain financial summary information, and fill the financial summary information into the corresponding position according to the template of the summary report, using simplified fonts and formats and icons and graphics. The report content should be concise and clear, and focus on key financial indicators and trends, such as pie charts to represent revenue share and line charts to represent profit trends. Review the summary version of the financial report by checking the correctness and readability of the charts and graphics to obtain the summary version of the financial report, ensure the accuracy of the data and the simplicity of the report, and output the summary version of the financial report in PDF format to ensure that the report can be easily viewed on mobile devices, and then deliver it to users to meet their needs for quick understanding of financial status; The process of generating financial statements for a specific department: According to the specific department specified by the user, the financial data related to the specific department is extracted from the financial database, the financial data related to the specific department is customized, and the data is screened and summarized according to the characteristics and needs of the specific department to obtain customized data. A template for the specific department version of the report is designed based on the key financial indicators and business characteristics of the specific department, and the customized data is filled into the template for the specific department version of the report to ensure that the report content is closely related to the department needs. The format is adjusted according to the style and requirements of the specific department. The person in charge of the specific department reviews the specific department version of the financial report to ensure the accuracy of the data and the practicality of the report, and gives review opinions. Adjustments and improvements are made based on the review opinions to obtain the specific department version of the financial report to ensure that the report meets the needs of the department. The specific department version of the financial report is output to the format used by the specific department and delivered to the designated department to provide strong financial support for the department's decision-making and management.

[0012] Preferably, the regulatory inspection module includes: The regulatory matching unit is used to match the generated financial statements with the built-in financial regulatory library to identify regulatory compliance in the statements. The financial regulatory library includes preset financial principles and financial standards. By comparing the financial statements with the financial regulatory library, the accuracy and completeness of the statements are checked to ensure quality control before the statements are released. The problem diagnosis unit is connected to the regulatory matching unit and is used to analyze the regulatory compliance generated by the regulatory matching unit, give improvement suggestions, and help users understand and solve regulatory compliance problems; The regulatory update unit is used to update the built-in financial regulatory library every hour to ensure that the platform's regulatory compliance checks are always based on the latest financial regulations and standards.

[0013] The present invention provides an intelligent financial statement automatic generation platform, which has the following beneficial effects: 1. The intelligent financial statement automatic generation platform can not only automatically classify and generate account attributions based on key financial characteristics and financial related accounts through the classification decision unit, but also supports users to customize the classification rules of financial accounts. This enables the platform to adapt to the specific financial reporting needs and business characteristics of different enterprises, and improves the flexibility and applicability of the system.

[0014] 2. The intelligent financial statement automatic generation platform calculates the summary amount of each account according to the financial structure data through the summary calculation unit, and generates report templates in combination with the enterprise financial report standards, including profit summary, asset-liability matching and cash liquidity analysis, providing an accurate data basis and standardized templates for the generation of financial statements, improving the professionalism and reliability of the reports, and through the report construction unit, it can build financial statements according to the report templates and report formats generated by the summary calculation unit, supporting multiple report formats and customized content; by adjusting the format of the report, beautifying it, inserting charts, annotations and performing ratio calculations, it improves the readability and information value of the report, meeting the needs of different users.

[0015] 3. The intelligent financial report automatic generation platform supports report customization through the platform and provides different versions of financial reports, including detailed version, summary version and specific department version, which meets the different needs of different users for financial information, improves the flexibility and applicability of the system, and is based on the customizable mechanism of classification rules. It can also customize the classification rules according to the actual needs or even temporary needs of users, so as to develop a temporary version exclusive to current needs, providing more targeted financial support for corporate decision-making and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The present invention is a flowchart of an intelligent financial statement automatic generation platform. DETAILED DESCRIPTION

[0017] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0018] like Figure 1As shown, the present invention provides a technical solution: an intelligent financial statement automatic generation platform, including a data integration module, which is used to connect and synchronize data sources, including bank transaction records, enterprise resource planning systems and electronic invoice systems, and to capture data by configuring the connection of data interfaces. The module supports data integration methods including API, direct database connection and file import; The identification and classification module is connected to the data integration module and is used to analyze and identify the correlation between the intermediate data and the financial subjects through machine learning algorithms, obtain the financial related subjects, and archive the subject attribution according to the preset financial subjects; The report generation module is connected with the identification and classification module and is used to generate financial reports by summarizing the classified accounts according to the enterprise financial reporting standards and the user-defined report format. It also supports report customization and provides different versions of financial reports. The regulatory verification module is connected to the report generation module and performs regulatory compliance checks on financial statements through the built-in financial regulatory library to verify whether the generated financial statements comply with the preset financial regulations and financial standards.

[0019] The data integration modules include: The interface configuration unit is used to set and manage the connection configuration of different data sources. The connection configuration includes authentication information, data capture frequency and data type. It also supports custom configuration to adapt to the data environment of different enterprises, improving the flexibility and convenience of data acquisition. A synchronization control unit is connected to the interface configuration unit and is used to control the synchronization process of the data source to ensure timely data update and synchronization accuracy, obtain log records by real-time monitoring of the state and performance of data source synchronization, and handle exceptions that occur during the synchronization process according to the log records; The data cache unit is connected to the synchronization control unit and is used to temporarily store the data synchronized from the data source, recorded as the original data, and manage the temporary storage and access of the data through an efficient cache strategy, thereby reducing the waiting time of the data processing module and improving the overall performance of the system; The data processing unit is connected to the data cache unit, receives the original data generated by the data cache unit, and cleans the original data, including removing duplicate records, format conversion, missing value processing and outlier detection, to obtain intermediate data in a unified format, and sends the intermediate data to the recognition and classification module.

[0020] The interface configuration unit is used to set and manage the connection configuration of different data sources as follows: For data sources that require authentication, collect the corresponding authentication information; for bank transaction records, collect usernames, passwords, digital certificates, dynamic passwords, corporate bank account information, and bank interface permission settings; for enterprise resource planning systems, collect the corresponding system usernames and passwords, user roles and permissions, database connection information, API keys, and the company's internal approval process and authorization documents; for electronic invoice systems, collect taxpayer identification numbers, usernames and passwords, electronic signatures, verification codes, and interface permission settings for the invoice platform; Determine the frequency of data capture based on business needs and the update frequency of the data source. When the data source updates frequently, set the frequency of data capture to once an hour. When the data source updates frequently, set the frequency of data capture to once a day. Analyze the format and structure of the data in the data source, determine the data type, and establish a connection with the data source based on the authentication information, data capture frequency, and data type. The process of custom configuration is as follows: Based on the connected data source, identify the business needs of the data source, analyze the data source with configuration requirements, obtain the characteristics of the data source, including data structure, update mechanism and authentication method, and determine how to effectively connect to and obtain data from the data source; design a custom configuration plan based on programming tools and programming techniques according to business needs and data source characteristics; the custom configuration plan includes authentication method, data capture frequency, data type processing and data filtering and screening; among them: Authentication method: When the authentication method of the data source is different from the common method, an authentication module is designed for authentication; Data crawling frequency: Determine the appropriate data crawling frequency based on the update frequency of the data source and the urgency of business needs; set a dynamically adjusted data crawling strategy to automatically adjust the crawling frequency based on changes in the data source; Data type processing: If the data type of the data source is special, design a corresponding data processing module to convert it into a standard format that the system can process; Data filtering and screening: Design data filtering and screening rules according to business needs to obtain only the required data and improve data processing efficiency; The process of getting the log record: The synchronization information of the data source that needs to be recorded is clarified. The synchronization information includes the data source connection information, the synchronization start time and end time, the number and progress of data capture, the performance indicators during the synchronization process, error information and warning information; wherein, the data source connection information includes the data source name, connection time and authentication information; the file log recording method is adopted, and the log recording logic is inserted into the code of the synchronization control unit. The log recording logic is: when connecting to the data source, the connection information and start time are recorded; during the data capture process, the number and progress of capture are recorded; at the end of the synchronization, the end time and performance indicators are recorded; when an error or warning occurs, the error information or warning information is recorded; and the log level is configured based on the synchronization information of the data source that needs to be recorded, thereby completing the log recording process; The process of handling exceptions that occur during synchronization based on log records Use scripts to scan log records regularly to find abnormal information, generate alarms based on abnormal information, obtain error information in log records based on alarms and perform abnormal analysis to obtain the cause of the abnormality, which includes: authentication failure, network failure, data format provided by the data source does not match system expectations, data transmission speed is too slow, response time is too long, code errors in the synchronization control unit itself, or other system component failures; take processing measures based on the cause of the abnormality, which include rechecking authentication information, repairing network failures, contacting data source administrators to solve problems, adjusting data processing logic, performing data format conversion or cleaning, optimizing data capture algorithms, increasing resource allocation, adjusting data capture frequency, repairing code errors in the synchronization control unit, restarting related system components, or performing system upgrades; obtain exception handling results based on the processing measures for the cause of the abnormality, and record the exception handling results in the log, which includes the measures taken, processing time, and result status, and completes the processing of exceptions that occur during the synchronization process based on log records, which is helpful for subsequent problem troubleshooting and system optimization.

[0021] The recognition and classification modules include: A feature extraction unit is used to identify and extract key financial features from the preprocessed intermediate data. The key financial features include amount, date, and event description, and use a machine learning algorithm to analyze the correlation between the key financial features and financial accounts to identify the financial related accounts corresponding to the data items; The classification decision unit is connected to the feature extraction unit and is used to classify the intermediate data into corresponding financial accounts according to the extracted key financial features and the obtained financial related accounts, and generate account attribution; it is also used to judge the account attribution of each record through the pre-trained classification model, obtain the judgment result, and support users to customize the classification rules of financial accounts to meet different financial reporting requirements; The data archiving unit is connected to the classification decision unit and is used to archive and store the classified accounts in the financial database by account, obtain financial structure data, ensure the orderliness and retrievability of the data, support efficient data query and update operations, and provide an accurate data source for report generation.

[0022] The pre-trained classification model determines the subject of each record and obtains the judgment result: The classification model is deployed to the classification decision unit of the recognition and classification module, and the prepared input data is input into the classification model for classification prediction, and the classification result is output. The classification result represents the probability that each record belongs to a different financial account, and the probability of each financial account is compared with the corresponding determined probability threshold to determine the account attribution judgment result of each record, wherein during the comparison: if the probability of the financial account is greater than or equal to the probability threshold, the record is classified into this financial account; if the probability of the financial account is less than the probability threshold, the record is marked for manual review; The process of building a classification model is: Clarify the classification of financial accounts, obtain account categories, which include assets, liabilities, equity, income and costs, clarify the characteristic indicators of classification, which include classification accuracy and recall rate, and use the characteristic indicators as the optimization target of the model, obtain financial data records based on the enterprise financial system and database and the public financial statement data set, the financial data records contain different types of financial transaction information, and have corresponding financial account annotations; clean the financial data records, remove outliers, erroneous data and duplicate data, obtain the financial data set, divide the financial data set into training set, validation set and test set, use the training set for model training and optimization, and use the validation set for training In the process, the model parameters are adjusted and the model performance is evaluated to prevent overfitting; the test set is used to evaluate the performance of the model on the data; according to the subject categories of the financial subject classification, a deep neural network is selected as the architecture of the classification model, the selected deep neural network is trained using the training set data, and the trained model is evaluated using the validation set data to obtain the evaluation results. According to the evaluation results, the parameters and hyperparameters of the model are adjusted to obtain the initial model, and the initial model is tested using the test set data to verify the performance of the model on unseen data, so that the verification results of the model meet the optimization goals, and finally the classification model is obtained, and the classification model is deployed to the classification decision unit of the recognition and classification module for classifying financial data.

[0023] The process of user-defined classification rules for financial accounts: First, clarify the financial reporting requirements, determine which specific financial account classifications are required and the classification standards; according to the financial reporting requirements, design customized financial account classification rules and define them based on financial characteristics: Transaction amount range: records are classified into different accounts according to the size of the transaction amount; for example, transactions with an amount less than 1,000 yuan are classified into the "small expenditure" account, and transactions with an amount greater than 10,000 yuan are classified into the "major expenditure" account; Transaction object type: Classify according to the nature of the transaction object; for example, transactions with suppliers are classified into the "Purchase Expenses" account, and transactions with customers are classified into the "Sales Revenue" account; Business type: classified according to the specific type of business; for example, R&D-related expenses are classified into the "R&D expenses" account, and sales-related expenses are classified into the "Sales expenses" account; Convert the designed classification rules into executable configuration files, and input the configuration files through the rule configuration interface provided by the system; test the custom classification rules with known financial data before application; check whether the rules can correctly classify the records into the expected subjects; if not, optimize and adjust the classification rules to obtain clear custom classification rules; then apply the tested classification rules to financial data processing; and in the subsequent use process, regularly check the validity of the classification rules, optimize and adjust them as the business changes and financial reporting requirements, and update and maintain the custom classification rules in a timely manner.

[0024] The process of regularly checking the effectiveness of classification rules and optimizing and adjusting them as business changes and financial reporting requirements change is as follows: Determine the inspection cycle of classification rules based on the business characteristics of the enterprise and the importance of financial reports. For example, for enterprises with rapidly changing businesses, inspections can be conducted monthly, and for enterprises with relatively stable businesses, inspections can be conducted every quarter or half a year. Collect feedback on the effectiveness of classification rules from three channels: financial personnel, management, and auditors. For financial personnel: they use classification rules to generate financial statements in their daily work and have an intuitive understanding of the actual effects of the rules. Their opinions and suggestions can be collected through questionnaires, interviews, etc.; Management: Management makes decisions based on financial statements and has high requirements for the accuracy and availability of financial data. Their feedback can help determine whether the classification rules meet decision-making needs; Auditors: Problems with classification rules may be discovered during the audit process. Communicate with auditors to understand their findings and suggestions; Check and analyze recent financial data, including checking the distribution of data, comparing data from different periods, and analyzing abnormal data to obtain financial data analysis results; Check the distribution of data: observe whether the data distribution under different financial accounts is in line with expectations. If the data under certain accounts is abnormally concentrated or dispersed, it means that there is a problem with the classification rules; Compare data from different periods: Compare financial data from different time periods to see whether the classification rules remain stable over time. If the classification results change significantly over time, further analysis of the reasons is needed; Analyze abnormal data: Pay attention to outliers and misclassified data to find out the reasons for these problems, whether the classification rules are not applicable to certain special cases, or there are problems with the data itself; Based on the collected feedback information and financial data analysis results, the effectiveness of the classification rules is evaluated to obtain the evaluation results. The evaluation content includes the evaluation of accuracy, consistency and adaptability. For accuracy: check the degree of conformity between the classification results and the actual business situation, and check a certain proportion of data through random sampling to see whether the classification is accurate; for consistency: ensure that different personnel use the same classification rules to obtain consistent results, conduct cross-validation, and let multiple financial personnel classify the same batch of data to compare the consistency of the results; for adaptability: evaluate whether the classification rules can adapt to business changes and new financial reporting requirements. If the business has undergone major changes, whether the classification rules can still effectively classify financial data. If problems are found in the classification rules, then optimize and adjust them according to the evaluation results. Optimization and adjustment include modifying the rules, retraining the model, and testing and verification; Among them, for modifying rules: according to the feedback information and analysis results, make specific modifications to the classification rules, for example, adjust the classification standards, add new classification subjects, and modify the scope of application of the rules; for retraining models: the model can be retrained with new data to improve the accuracy and adaptability of classification; for testing and verification: Before implementing new classification rules, sufficient testing and verification can be carried out. A portion of the data can be used for simulated classification to check the effectiveness of the new rules. If necessary, relevant personnel can be invited to try them out and collect their feedback. Subsequently, the optimized classification rules can be determined, and training materials and guidelines can be generated based on the optimized classification rules. Communicate and train financial personnel and management to help relevant personnel adapt to the new classification method by providing training materials and guidance. New rules and changes can be introduced in a timely manner to ensure that relevant personnel understand and can correctly apply the new rules. After the classification rules are adjusted, the optimized classification rules and their effects should be continuously monitored, feedback information should be collected regularly, financial data should be analyzed, the effectiveness of the classification rules should continue to be evaluated, and adjustments and optimizations should be made in a timely manner in combination with business changes and new needs to obtain new classification rules to ensure that the classification rules can always meet the company's financial reporting needs.

[0025] The report generation module includes: The summary calculation unit is used to calculate the summary amount of each account based on the financial structure data, and generate a report template in combination with the enterprise financial report standard. The report template includes profit summary, asset-liability matching and cash liquidity analysis; The report construction unit is connected to the summary calculation unit and is used to construct financial reports based on the report template and report format generated by the summary calculation unit. It supports multiple report formats and customized content, chart insertion, annotation addition and ratio calculation to improve the readability and information value of the report.

[0026] The process of constructing financial statements by the report construction unit: Parse the report template, clarify the report structure, report format and data fields required for the report. For example, the profit summary report requires income, cost and expenses, the asset-liability matching report requires assets, liabilities and owner's equity, and the cash flow analysis report requires cash inflows and cash outflows. Identify the formulas and calculation logic in the report. For example, the net profit calculation formula in the profit summary report is income minus costs and expenses, and the debt-to-asset ratio calculation formula in the asset-liability matching report is total liabilities divided by total assets. According to the requirements of the report template, extract the corresponding financial account data from the financial structure data to obtain the extracted information. For example, if the report template requires the data of the income account, find the amount data corresponding to the income account from the financial structure data. For the data fields that need to be calculated, calculate according to the formulas and calculation logic in the report template to obtain the calculated data. For example, when calculating the net profit, substitute the income, cost and expense data into the net profit calculation formula for calculation. Fill the extracted information and calculated data into the corresponding positions of the report template. Format and layout the title, subtitle and notes in the report to make them clear. Easy to read. The format of the report is adjusted and beautified according to the user's requirements and aesthetic standards. For example, the format attributes such as font, font size, color, border, etc. are set to make the report more beautiful and professional. The report layout is adjusted using charts and graphics to ensure that the data is presented reasonably and clearly, and the readability and intuitiveness of the report are enhanced to obtain the financial report. The constructed financial report is then manually reviewed and verified to check the accuracy of the data, the correctness of the formula and the completeness of the report. If errors or problems are found, they are corrected and adjusted in a timely manner to ensure the quality and reliability of the financial report. After the financial report is reviewed and verified, it is output to the format required by the user, such as PDF, Excel, HTML, to ensure that the report can be viewed and printed on different devices and software, and support report customization to provide different versions of financial reports according to user needs. Different versions of financial reports include detailed version, summary version and specific department version to meet the needs of different users. Through the above process, the report construction unit can construct accurate, beautiful and professional financial reports based on the report template and financial structure data generated by the summary calculation unit.

[0027] Based on report customization, the process of providing different versions of financial statements according to user needs is as follows: Detailed financial statement generation process: Obtain financial structure data, arrange the financial structure data according to the requirements of financial statements, organize and classify the financial structure data, fill the organized financial structure data one by one into the corresponding positions according to the template of the detailed report, ensure that each financial account has a detailed display, carry out calculations for the fields that need to be calculated, obtain the calculation results, and display the calculation process and results in the report in the form of annotations. In this way, some complex financial accounts or transactions can be explained so that users can better understand the report content, set the font, font size and color to make the report easy to read, and use different colors to distinguish different financial accounts and important information, and add bar charts and line charts to show the trends and changes of financial data, adjust the page layout, ensure the overall appearance and professionalism of the report, and invite auditors to review and verify the detailed financial statements. The review and verification includes checking the accuracy of the data, the correctness of the calculation and the standardization of the format to ensure that the report meets the financial specifications and requirements. Then, the approved detailed financial statements are output based on the format required by the user, and the reports are sent to the users via email or file sharing platforms to complete the delivery. Summary financial statement generation process: Obtain the key financial features generated in the feature extraction unit, summarize and calculate the key financial features, and obtain financial summary information. For example, calculate the revenue share and expense growth rate of different business segments. According to the template of the summary report, use simplified fonts and formats and fill the financial summary information in the corresponding positions through icons and graphics. The report content should be concise and clear, and focus on key financial indicators and trends. For example, pie charts represent revenue share and line charts represent profit trends. Use simple fonts and formats to make the report easy to read quickly and avoid excessive decoration and complex layout. At the same time, ensure that the charts and graphics are clear and readable, and the color matching is coordinated to enhance the visualization of the report. Review the summary financial report by checking the correctness and readability of the charts and graphics to obtain the summary financial report, ensure the accuracy of the data and the simplicity of the report, and output the summary financial report in PDF format to ensure that the report can be easily viewed on mobile devices. Then deliver it to users to meet users' needs for quick understanding of financial status. The process of generating financial statements for a specific department: According to the specific department specified by the user, the financial data related to the specific department is extracted from the financial database. For example, if it is a report of the sales department, relevant data such as sales revenue and sales expenses are extracted. The financial data related to the specific department is customized, and filtered and summarized according to the characteristics and needs of the specific department to obtain customized data. For example, for the sales department, specific indicators such as sales gross profit margin and sales growth rate can be calculated. The template of the report of the specific department is designed based on the key financial indicators and business characteristics of the specific department. For example, the report of the sales department can focus on displaying sales performance, customer distribution and other information. The customized data is filled into the template of the report of the specific department to ensure that the report content is consistent with the department's needs. Closely related, the format is adjusted according to the style and requirements of a specific department to make the report conform to the image and culture of the department. For example, the report of the sales department can use lively colors and charts to reflect the vitality of the sales team, ensure the readability and visualization of the report, and facilitate department personnel to quickly understand the financial data. The head of a specific department will review the financial statements of the specific department to ensure the accuracy of the data and the practicality of the report, give review opinions, make adjustments and improvements based on the review opinions, obtain the financial statements of the specific department, ensure that the report meets the needs of the department, output the financial statements of the specific department to the format used by the specific department, and deliver them to the designated department to provide strong financial support for the department's decision-making and management.

[0028] Regulatory inspection modules include: The regulatory matching unit is used to match the generated financial statements with the built-in financial regulatory library to identify regulatory compliance in the statements. The financial regulatory library includes preset financial principles and financial standards. By comparing the financial statements with the financial regulatory library, the accuracy and completeness of the statements are checked to ensure quality control before the statements are released. The problem diagnosis unit is connected to the regulatory matching unit and is used to analyze the regulatory compliance generated by the regulatory matching unit, give improvement suggestions, and help users understand and solve regulatory compliance problems; The regulatory update unit is used to update the built-in financial regulatory library every hour to ensure that the platform's regulatory compliance checks are always based on the latest financial regulations and standards.

[0029] It should be further explained that, in the specific implementation process, the management of the data source connection configuration and the control of the synchronization process are realized through the interface configuration unit and the synchronization control unit; the timely updating, accuracy and synchronization stability of the data are ensured, and at the same time, the maintainability and reliability of the system are improved through the logging and exception handling mechanism; the existence of the data cache unit further improves the system performance and reduces the data processing waiting time.

[0030] The feature extraction unit can extract key financial features from intermediate data, and use machine learning algorithms to analyze the correlation between these features and financial subjects, identify the financial related subjects corresponding to the data items, improve the accuracy and automation of financial subject classification, and reduce the errors and workload of manual classification; the data archiving unit can archive and store the classified subjects in the financial database according to the subjects, ensuring the orderliness and retrievability of the data; provide an accurate data source for report generation, facilitate subsequent data query and update operations, and improve the management efficiency of financial data.

[0031] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field and related fields without creative work should fall within the scope of protection of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention are implemented according to the conventional means in the field unless otherwise specified and limited.

Claims

1. An intelligent financial statement automatic generation platform, characterized by: include: Data integration module, used to connect and synchronize data sources, including bank transaction records, enterprise resource planning systems and electronic invoice systems, and to capture data by configuring the connection of data interfaces. This module supports data integration methods including API, direct database connection and file import; The identification and classification module is connected to the data integration module and is used to analyze and identify the correlation between the intermediate data and the financial subjects through machine learning algorithms, obtain the financial related subjects, and archive the subject attribution according to the preset financial subjects; The report generation module is connected with the identification and classification module and is used to generate financial reports by summarizing the classified accounts according to the enterprise financial reporting standards and the user-defined report format. It also supports report customization and provides different versions of financial reports. The regulatory verification module is connected to the report generation module and performs regulatory compliance checks on financial statements through the built-in financial regulatory library to verify whether the generated financial statements comply with the preset financial regulations and financial standards.

2. The intelligent financial statement automatic generation platform according to claim 1, characterized in that: The data integration module includes: The interface configuration unit is used to set and manage the connection configuration of different data sources. The connection configuration includes authentication information, data capture frequency and data type, and also supports custom configuration; A synchronization control unit is connected to the interface configuration unit and is used to control the synchronization process of the data source, obtain log records by real-time monitoring of the state and performance of the data source synchronization, and handle exceptions that occur during the synchronization process according to the log records; A data cache unit, connected to the synchronization control unit, for temporarily storing data synchronized from the data source, recorded as original data; The data processing unit is connected to the data cache unit, receives the original data generated by the data cache unit, and cleans the original data, including removing duplicate records, format conversion, missing value processing and outlier detection, to obtain intermediate data in a unified format, and sends the intermediate data to the recognition and classification module.

3. The intelligent financial statement automatic generation platform according to claim 2 is characterized by: The identification and classification module comprises: A feature extraction unit is used to identify and extract key financial features from the preprocessed intermediate data. The key financial features include amount, date, and event description, and use a machine learning algorithm to analyze the correlation between the key financial features and financial accounts to identify the financial related accounts corresponding to the data items; The classification decision unit is connected to the feature extraction unit and is used to classify the intermediate data into corresponding financial accounts according to the extracted key financial features and the obtained financial related accounts, and generate account attribution; it is also used to judge the account attribution of each record through the pre-trained classification model, obtain the judgment result, and support users to customize the classification rules of financial accounts to meet different financial reporting requirements; The data archiving unit is connected to the classification decision unit and is used to archive and store the classified accounts in the financial database according to the accounts to obtain the financial structure data.

4. The intelligent financial statement automatic generation platform according to claim 3 is characterized by: The pre-trained classification model determines the subject attribution of each record and obtains the judgment result: The classification model is deployed to the classification decision unit of the recognition and classification module, and the prepared input data is input into the classification model for classification prediction, and the classification result is output. The classification result represents the probability that each record belongs to a different financial account, and the probability of each financial account is compared with the corresponding determined probability threshold to determine the account attribution judgment result of each record. During the comparison: if the probability of the financial account is greater than or equal to the probability threshold, the record is classified into this financial account; if the probability of the financial account is less than the probability threshold, the record is marked for manual review.

5. The intelligent financial statement automatic generation platform according to claim 4 is characterized by: The process of user-defined classification rules for financial accounts: First, clarify the financial reporting requirements, design customized classification rules for financial accounts based on the financial reporting requirements, and define them based on financial characteristics: Transaction amount range: records are classified into different accounts according to the size of the transaction amount; Transaction object type: classified according to the nature of the transaction object; Business type: classified according to the specific type of business; Convert the designed classification rules into executable configuration files, and input the configuration files through the rule configuration interface provided by the system; test the custom classification rules with known financial data before application; check whether the rules can correctly classify the records into the expected subjects; if not, optimize and adjust the classification rules to obtain clear custom classification rules; then apply the tested classification rules to financial data processing; and in the subsequent use process, regularly check the effectiveness of the classification rules, and optimize and adjust them as the business changes and financial reporting requirements.

6. The intelligent financial statement automatic generation platform according to claim 5 is characterized by: The process of regularly checking the effectiveness of classification rules and optimizing and adjusting them as business changes and financial reporting requirements change is as follows: According to the business characteristics of the enterprise and the importance of financial reports, the inspection cycle of classification rules is determined. Feedback information on the effectiveness of classification rules is collected from three channels: financial personnel, management and auditors. Recent financial data is inspected and analyzed. The inspection and analysis includes checking the distribution of data, comparing data from different periods and analyzing abnormal data to obtain financial data analysis results. The effectiveness of classification rules is evaluated based on the collected feedback information and financial data analysis results to obtain evaluation results. The evaluation content includes accuracy, consistency and adaptability. If problems are found in the classification rules, optimization and adjustment are made according to the evaluation results. Optimization and adjustment include modifying the rules, retraining the model, and testing and verification. Subsequently, the optimized classification rules are determined, and training materials and guidelines are generated based on the optimized classification rules. Communicate and train financial personnel and management. After the classification rules are adjusted, the optimized classification rules and their effects are continuously monitored. Feedback information is collected regularly, financial data is analyzed, and the effectiveness of classification rules continues to be evaluated. In combination with business changes and new needs, timely adjustments and optimizations are made to obtain new classification rules.

7. The intelligent financial statement automatic generation platform according to claim 6, characterized in that: The report generation module includes: The summary calculation unit is used to calculate the summary amount of each account based on the financial structure data, and generate a report template in combination with the enterprise financial report standard. The report template includes profit summary, asset-liability matching and cash liquidity analysis; The report building unit is connected with the summary calculation unit and is used to build a financial report according to the report template and report format generated by the summary calculation unit.

8. The intelligent financial statement automatic generation platform according to claim 7, characterized in that: The process of constructing financial statements by the report construction unit: Parse the report template, clarify the report structure, report format and data fields required for the report, identify the formulas and calculation logic in the report, extract the corresponding financial account data from the financial structure data according to the requirements of the report template, obtain the extracted information, and for the data fields that need to be calculated, calculate according to the formulas and calculation logic in the report template to obtain the calculated data, and fill the extracted information and calculated data into the corresponding positions of the report template, wherein the title, subtitle and comments in the report are formatted and laid out, and the format of the report is adjusted and beautified according to the user's requirements and aesthetic standards, and the report layout is adjusted with charts and graphics to obtain the financial report, and then the constructed financial report is manually reviewed and verified. If errors or problems are found, they are corrected and adjusted in a timely manner. After the financial report passes the review and verification, it is output to the format required by the user, and report customization is supported to provide different versions of financial reports according to user needs. Different versions of financial reports include detailed version, summary version and specific department version.

9. The intelligent financial statement automatic generation platform according to claim 8, characterized in that: Based on report customization, the process of providing different versions of financial statements according to user needs is as follows: Detailed financial statement generation process: Obtain financial structure data, arrange the financial structure data according to the requirements of financial statements, organize and classify the financial structure data, fill the organized financial structure data into the corresponding positions one by one according to the template of the detailed report, calculate the fields that need to be calculated, obtain the calculation results, and display the calculation process and results in the report in the form of annotations, set the font, font size and color, and add bar charts and line charts to show the trend and changes of financial data, adjust the page layout, invite auditors to review and verify the detailed financial statements, the review and verification includes checking the accuracy of data, the correctness of calculations and the standardization of formats, and then output the approved detailed financial statements based on the format required by the user, and send the report to the user by email to complete the delivery. Summary financial statement generation process: Obtain the key financial features generated in the feature extraction unit, summarize and calculate the key financial features to obtain financial summary information, fill the financial summary information into corresponding positions according to the template of the summary report, use simplified fonts and formats, and use icons and graphics to review the summary financial report by checking the correctness and readability of the charts and graphics to obtain the summary financial report, output the summary financial report into PDF format, and then deliver it to the user; The process of generating financial statements for a specific department: According to the specific department specified by the user, the financial data related to the specific department is extracted from the financial database, the financial data related to the specific department is customized, and the data is screened and summarized according to the characteristics and needs of the specific department to obtain customized data; a template for the specific department version report is designed based on the key financial indicators and business characteristics of the specific department, the customized data is filled into the template for the specific department version report, and the format is adjusted according to the style and requirements of the specific department. The person in charge of the specific department reviews the specific department version of the financial report and gives review opinions. Adjustments and improvements are made based on the review opinions to obtain the specific department version of the financial report, which is output to the format used by the specific department and delivered to the designated department.

10. The intelligent financial statement automatic generation platform according to claim 9, characterized in that: The regulatory inspection module includes: The regulatory matching unit is used to match the generated financial statements with the built-in financial regulatory library to identify regulatory compliance in the statements. The financial regulatory library includes preset financial rules and financial standards. A problem diagnosis unit, connected to the regulation matching unit, is used to analyze the regulatory compliance generated by the regulation matching unit and provide improvement suggestions; The regulations update unit is used to update the built-in financial regulations library every hour.

Images