Automatic auditing management system for intelligent financial statements

Abstract

The invention discloses an automatic auditing management system for an intelligent financial statement, relates to the technical field of financial auditing, and solves the problems that in the actual financial statement auditing process, when a financial statement image is converted into an electronic financial statement, errors occur in the conversion process due to the fact that the financial statement image is not clear, and the financial statement image is not clear. And the technical problem of financial statement auditing is influenced. Obtaining a financial statement image uploaded by a user; generating an electronic financial statement according to the financial statement image; analyzing the conversion accuracy according to the electronic financial statement; obtaining risk data of the electronic financial statement; calculating a risk coefficient of the electronic financial statement according to the risk data; determining an auditing level of the electronic financial statement based on the risk coefficient; auditing the electronic financial statement according to the auditing level; dividing the audited electronic financial statement to obtain the grade of the electronic financial statement; encrypting and storing the electronic financial statement according to the grade of the electronic financial statement; and the auditing efficiency and quality can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of financial audit, and relates to the audit management technology of financial statements. Specifically, it is an automatic audit management system for intelligent financial statements. Background Art

[0002] Financial statements are important documents for enterprises to reflect their financial status, operating results and cash flows, usually including balance sheets, income statements, cash flow statements and statements of changes in shareholders' equity, etc.; intelligent automatic audit of financial statements can quickly process a large amount of financial data, and use preset rules and algorithms for automated audit, significantly reducing the time of manual audit; automatic audit can reduce the possibility of human errors through automated data processing and audit processes; the system can automatically identify, classify and process various financial data to ensure the accuracy and consistency of the data; intelligent automatic audit can quickly identify potential risk points through real-time monitoring and analysis of financial data, and take corresponding early warning and control measures in a timely manner; it helps to reduce the financial risks of enterprises and ensure the healthy development of enterprises; automatic audit records and tracks the approval process of financial statements to ensure that each link is traceable, improving the transparency of the approval process.

[0003] Existing financial statement audit systems use set rules and logics to audit uploaded electronic financial statements; when the financial statement uploaded by the user is an image, the image of the financial statement is converted into an electronic financial statement, and then the electronic financial statement is audited; however, in actual financial statement audits, when converting the financial statement image into an electronic financial statement, errors may occur during the conversion due to unclear financial statement images, affecting the audit of financial statements.

[0004] The present invention provides an automatic audit management system for intelligent financial statements to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; for this purpose, the present invention proposes an automatic audit management system for intelligent financial statements, which is used to solve the technical problem that in actual financial statement audits, when converting the financial statement image into an electronic financial statement, errors may occur during the conversion due to unclear financial statement images, affecting the audit of financial statements.

[0006] To achieve the above object, the first aspect of the present invention provides an automatic audit management system for intelligent financial statements, including: an automatic audit module, and a data collection module and a terminal management module connected thereto;

[0007] Data acquisition module: used to obtain the financial statement images uploaded by users; generate electronic financial statements based on the financial statement images; analyze the conversion accuracy rate according to the electronic financial statements;

[0008] Automatic audit module: used to obtain the risk data of the electronic financial statements; calculate the risk coefficient of the electronic financial statements according to the risk data; determine the audit level of the electronic financial statements based on the risk coefficient; audit the electronic financial statements according to the audit level;

[0009] Terminal management module: used to classify the electronic financial statements that have passed the audit to obtain the levels of the electronic financial statements; encrypt and store the electronic financial statements according to the levels of the electronic financial statements.

[0010] Preferably, the generation of the electronic financial statements based on the financial statement images includes:

[0011] Retrieve the financial statement images and preprocess the financial statement images; wherein, the preprocessing includes: cropping, denoising, brightness adjustment, and skew correction;

[0012] Import the financial statement images into the OCR character recognition software to generate electronic financial statements; wherein, the OCR character recognition software includes: Rossum or Klippa.

[0013] It should be noted that the OCR character recognition software converts the text information in the paper document or image into digital signals through optical scanning, and then uses the character recognition algorithm to analyze and process these digital signals, so as to recognize the text content therein and convert it into an editable electronic text format.

[0014] The present invention preprocesses the financial statement images, adjusts the images to the most suitable size and brightness, which is convenient for image recognition; uses the OCR character recognition software to convert the financial statement images into electronic financial statements, which can facilitate the intelligent processing of financial statements.

[0015] Preferably, the analysis of the conversion accuracy rate according to the electronic financial statements includes:

[0016] Retrieve the electronic financial statements and the financial statement images; use the functions in Python to convert the electronic financial statements into images to obtain the converted images;

[0017] Compare the converted images with the financial statement images to generate the conversion similarity; mark the conversion similarity as the conversion accuracy rate of the electronic financial statements.

[0018] The present invention reconverts an electronic financial statement into an image to obtain a converted image, compares the converted image with a financial statement image to obtain a conversion similarity, and marks the conversion similarity as the conversion accuracy rate of the electronic financial statement, which can evaluate the steps of converting a financial statement image into an electronic financial statement and lay a foundation for subsequent risk assessment of the electronic financial statement.

[0019] Preferably, the method for obtaining the conversion similarity includes:

[0020] Retrieve the converted image and the financial statement image, crop the financial statement image and the converted image into the same size, and divide the financial statement image and the converted image into several comparison regions according to the set region size.

[0021] Match the comparison regions of the financial statement image and the converted image to obtain several groups of comparison images, and integrate the several groups of comparison images into a comparison input sequence.

[0022] Call a similarity analysis model, input the comparison input sequence into the similarity analysis model to obtain the conversion similarity between the financial statement image and the converted image, where the similarity analysis model is constructed based on an artificial intelligence model.

[0023] It should be noted that when cropping the converted image and the financial statement image, only the larger-sized image needs to be cropped, and the image can be appropriately enlarged or reduced during cropping.

[0024] The present invention crops the converted image and the financial statement image into the same size, divides and matches the financial statement image and the converted image to obtain several groups of comparison images, analyzes the comparison images using a similarity analysis model to obtain the conversion similarity between the two, can evaluate the conversion accuracy rate of the financial statement image, is conducive to improving the accuracy of the evaluation result, and provides data support for subsequent risk analysis of the financial statement.

[0025] Preferably, the similarity analysis model is constructed based on an artificial intelligence model and includes:

[0026] Obtain a standard data set, where the standard data set includes standard input data consistent with the content attributes of the comparison input sequence and standard input data consistent with the content attributes of the conversion similarity.

[0027] Divide the standard data set into a training set, a validation set, and a test set according to a set ratio, train the artificial intelligence model using the training set, adjust the internal parameters of the artificial intelligence model using the validation set, and test the artificial intelligence model using the test set to obtain test metrics.

[0028] Determine whether the test metrics are greater than the metric threshold. If so, mark the trained artificial intelligence model as a similarity analysis model; if not, retrain the artificial intelligence model. Among them, the artificial intelligence model includes: a convolutional neural network model or a long short-term memory neural network model.

[0029] It should be noted that the test metrics include accuracy, recall, stability, and F1 score; the metric threshold is obtained through experimental simulation; the division ratio of the standard data set is set by expert evaluation. When the artificial intelligence model needs to be retrained, the division ratio of the standard data set is re-evaluated.

[0030] Preferably, calculating the risk coefficient of the electronic financial statement based on the risk data includes:

[0031] Retrieve the risk data of the electronic financial statement. Among them, the risk data includes: conversion accuracy rate, statement amount, and the creditworthiness of the reimbursement personnel;

[0032] Mark the conversion accuracy rate, statement amount, and the creditworthiness of the reimbursement personnel as ZQ, BJ, and BX; through the formula Calculate the risk coefficient of the electronic financial statement. Among them, α and β are proportionality coefficients greater than 0; SJ represents the set reimbursement amount threshold; DL is the dimension removal coefficient.

[0033] It should be noted that the creditworthiness of the reimbursement personnel is analyzed based on the historical data of the reimbursement personnel; the creditworthiness of the reimbursement personnel is updated regularly; the proportionality coefficient is set according to the actual situation; the reimbursement amount threshold is set according to the historical data, and the dimension removal coefficient is used to remove the dimension in the formula, generally set to 1.

[0034] The present invention calculates the risk coefficient of the electronic financial statement based on the risk data of the electronic financial statement, which can evaluate the risk of the electronic financial statement and provide data support for subsequent analysis of the review level of the electronic financial statement.

[0035] Preferably, determining the review level of the electronic financial statement based on the risk coefficient includes:

[0036] Retrieve the risk coefficient of the electronic financial statement and obtain the coefficient threshold; compare the risk coefficient with the coefficient threshold respectively. Among them, the coefficient threshold includes a first-level threshold and a second-level threshold;

[0037] Judge whether the risk coefficient is greater than the first-level threshold. If so, compare the risk coefficient with the second-level threshold; if not, set the review level of the corresponding electronic financial statement to the first level;

[0038] Determine whether the risk coefficient is greater than the secondary threshold; if yes, set the audit level of the corresponding electronic financial statement to level three; if no, set the audit level of the corresponding electronic financial statement to level two.

[0039] It should be noted that the audit levels, from smallest to largest, include: level one, level two, and level three; different audit levels of electronic financial statements correspond to different audit processes; the specific audit processes are set according to the actual situation.

[0040] The present invention analyzes the audit level of an electronic financial statement based on the risk coefficient and coefficient threshold of the electronic financial statement; it can conduct audits according to the actual evaluation of the electronic financial statement, strengthen the audit of electronic financial statements with high risks, and quickly audit electronic financial statements with low risks, which is conducive to improving the efficiency and quality of the audit.

[0041] Preferably, the auditing of the electronic financial statement according to the audit level includes:

[0042] Retrieve the audit level corresponding to the electronic financial statement, and match the corresponding audit process according to the audit level of the electronic financial statement;

[0043] When the audit level of the electronic financial statement is level one, conduct a preliminary audit on the electronic financial statement; when the audit level of the electronic financial statement is level two, after conducting a preliminary audit on the electronic financial statement, send it to the financial personnel, and the financial personnel shall confirm the funds of the electronic statement;

[0044] When the audit level of the electronic financial statement is level three, send the electronic financial statement after the preliminary audit and the audit by the financial personnel to the final auditor, and the final auditor shall audit the electronic financial statement;

[0045] Mark the electronic financial statement that has passed the audit as audited, return the electronic financial statement that has not passed the audit, and notify the corresponding reimbursement personnel to verify and then re-upload the financial statement image.

[0046] The present invention conducts corresponding audits on electronic financial statements according to the audit levels of the electronic financial statements, can strengthen the audit of electronic financial statements with good audit levels, and set multiple audit processes, which can reduce the situation of audit errors and is conducive to improving the accuracy of audit results.

[0047] Preferably, the division of the electronic financial statement that has passed the audit to obtain the level of the electronic financial statement includes:

[0048] Obtain the information of the reimbursement personnel of the electronic financial statement that has passed the audit; among them, the information of the reimbursement personnel includes the department of the reimbursement personnel and the level of the reimbursement personnel;

[0049] The electronic financial statements are numbered according to the departments of the reimbursement personnel to obtain large groups of electronic financial statements; the electronic financial statements in each large group are divided according to the levels of the reimbursement personnel to obtain the levels of the electronic financial statements.

[0050] It should be noted that the levels of the reimbursement personnel are set according to their positions. When the reimbursement personnel are promoted or transferred, their levels are re-determined according to the new positions; the numbering according to the departments of the personnel is represented by capital English letters.

[0051] The present invention classifies the electronic financial statements according to the departments and levels of the reimbursement personnel, which is beneficial to the management and query of the electronic financial statements.

[0052] Preferably, the encryption storage of the electronic financial statements according to their levels includes:

[0053] Retrieve the levels of the electronic financial statements; obtain the level encryption library; match the corresponding level encryption library according to the levels of the electronic financial statements, and randomly select an encryption algorithm from the level encryption library to encrypt the corresponding electronic financial statements;

[0054] Construct a corresponding department blockchain according to the levels of the electronic financial statements, store the encrypted electronic financial statements in the corresponding department blockchain according to the numbers; integrate several department blockchains to obtain a financial statement blockchain.

[0055] It should be noted that the level encryption library is set according to the encryption degree of the encryption algorithm, and each level of electronic financial statements corresponds to a level encryption library.

[0056] The present invention selects the corresponding encryption algorithm according to the levels of the electronic financial statements and stores the encrypted electronic financial statements in the corresponding blockchains; it can prevent the financial statement data from being maliciously tampered with, which is beneficial to ensuring the data security and privacy of the financial statements.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] 1. The present invention pre-processes the financial statement image, adjusts the image to the most appropriate size and brightness, and facilitates the recognition of the image; uses OCR text recognition software to convert the financial statement image into an electronic financial statement, which can facilitate the intelligent processing of the financial statement; converts the electronic financial statement into an image to obtain a converted image; compares the converted image with the financial statement image to obtain the conversion similarity; marks the conversion similarity as the conversion accuracy of the electronic financial statement; can evaluate the steps of converting the financial statement image into the electronic financial statement, laying a foundation for the subsequent risk assessment of the electronic financial statement; cuts the conversion image and the financial statement image into the same size, divides and matches the financial statement image and the conversion image, and obtains several groups of comparison images; uses a similarity analysis model to analyze the comparison image to obtain the conversion similarity between the two; can evaluate the conversion accuracy of the financial statement image, which is conducive to improving the accuracy of the evaluation result and provides data support for the subsequent analysis of the risk of the financial statement; calculates the risk coefficient of the electronic financial statement based on the risk data of the electronic financial statement, can evaluate the risk of the electronic financial statement, and provides data support for the subsequent analysis of the audit level of the electronic financial statement.

[0059] 2. The present invention analyzes the audit level of electronic financial statements according to the risk coefficient and coefficient threshold of the electronic financial statements; can conduct audits according to the actual evaluation of the electronic financial statements, strengthen the audit of electronic financial statements with high risks, and quickly audit electronic financial statements with low risks, which is conducive to improving the efficiency and quality of the audit; conduct corresponding audits on electronic financial statements according to the audit level of the electronic financial statements, which can strengthen the audit of electronic financial statements with good audit levels, and set up multiple audit processes, which can reduce the situation of audit errors and help improve the accuracy of the audit results; classify electronic financial statements according to the department and level of the account reporting personnel, which is conducive to the management and query of electronic financial statements; select the corresponding encryption algorithm according to the level of the electronic financial statement, and store the encrypted electronic financial statements according to the corresponding blockchain; can prevent the financial statement data from being maliciously tampered with, which is conducive to ensuring the data security and privacy of the financial statements. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 It is a schematic diagram of the steps of the overall scheme of the present invention;

[0062] Figure 2 It is a schematic diagram of the risk assessment steps of the financial statements of the present invention;

[0063] Figure 3 It is a schematic diagram of the audit management steps of the financial statements of the present invention. Specific embodiments

[0064] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0065] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides an automatic audit management system for intelligent financial statements, including: an automatic audit module, and a data collection module and a terminal management module connected thereto;

[0066] Data collection module: used to obtain the financial statement images uploaded by users; generate electronic financial statements according to the financial statement images; analyze and convert the accuracy rate according to the electronic financial statements;

[0067] Automatic audit module: used to obtain the risk data of the electronic financial statements; calculate the risk coefficient of the electronic financial statements according to the risk data; determine the audit level of the electronic financial statements based on the risk coefficient; audit the electronic financial statements according to the audit level;

[0068] Terminal management module: used to divide the electronic financial statements that have passed the audit to obtain the levels of the electronic financial statements; encrypt and store the electronic financial statements according to the levels of the electronic financial statements.

[0069] Please refer to Figure 2 , obtain the financial statement images uploaded by users; preprocess the financial statement images; wherein, the preprocessing includes: cropping, denoising, brightness adjustment, and correcting inclination; import the financial statement images into the OCR character recognition software to generate electronic financial statements; wherein, the OCR character recognition software includes: Rossum or Klippa;

[0070] Use the function in Python to convert the electronic financial statement into an image to obtain a converted image; crop the financial statement image and the converted image to the same size; divide the financial statement image and the converted image into several comparison areas according to the set area size; match the comparison areas of the financial statement image and the converted image to obtain several groups of comparison images; integrate the several groups of comparison images into a comparison input sequence; call the similarity analysis model, input the comparison input sequence into the similarity analysis model, and obtain the conversion similarity between the financial statement image and the converted image; mark the conversion similarity as the conversion accuracy of the electronic financial statement.

[0071] Obtain risk data of electronic financial statements, including: conversion accuracy, report amount, and credit rating of accountants; mark conversion accuracy, report amount, and credit rating of accountants as ZQ, BJ, and BX; and use the formula Calculate the risk factor of the electronic financial statement; where α and β are proportional coefficients greater than 0; SJ represents the set reimbursement amount threshold; DL is the dimension removal coefficient.

[0072] It should be noted that when the conversion accuracy or the credit rating of the account reporting personnel is higher, the e BX+ZQ The larger the value of -e BX+ZQ The smaller the value of The smaller the value of , the smaller the risk coefficient is. At this time, the smaller the value of the risk coefficient is, the risk coefficient is negatively correlated with the conversion accuracy and the credit of the reimbursement personnel. When the reimbursement amount is larger, the value of β×(BJ-SJ) is larger. At this time, the larger the value of the risk coefficient is, the risk coefficient is positively correlated with the reimbursement amount. When the value of β×(BJ-SJ) is negative, the default value of β×(BJ-SJ) is 0.

[0073] It is worth noting that the similarity analysis model is built based on an artificial intelligence model, including:

[0074] Acquire a standard data set; wherein the standard data set includes standard input data consistent with the content attributes of the comparison input sequence and standard input data consistent with the content attributes of the conversion similarity;

[0075] Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence model; use the test set to test the artificial intelligence model and obtain the test indicators;

[0076] Determine whether the test index is greater than the index threshold. If yes, mark the trained artificial intelligence model as a similarity analysis model; if no, retrain the artificial intelligence model; wherein the artificial intelligence model includes: a convolutional neural network model or a long short-term memory neural network model.

[0077] It should be noted that converting the financial statement image into an electronic financial statement, and then converting the electronic financial statement into a converted image. By comparing the conversion similarity between the converted image and the financial statement image, the conversion accuracy rate of the electronic financial statement is obtained, providing a data basis for analyzing the risks of the electronic financial statement; calculating the risk coefficient of the electronic financial statement based on the risk data of the electronic financial statement is conducive to multiple audits of electronic financial statements with high risks and improving the audit efficiency.

[0078] Please refer to Figure 3 , retrieve the risk coefficient of the electronic financial statement to obtain the coefficient threshold; compare the risk coefficient with the coefficient threshold respectively; among them, the coefficient threshold includes a primary threshold and a secondary threshold; determine whether the risk coefficient is greater than the primary threshold; if so, compare the risk coefficient with the secondary threshold; if not, set the audit level of the corresponding electronic financial statement to level one; determine whether the risk coefficient is greater than the secondary threshold; if so, set the audit level of the corresponding electronic financial statement to level three; if not, set the audit level of the corresponding electronic financial statement to level two.

[0079] For example: Suppose there are three electronic financial statements that need to be audited. After calculation, the risk coefficients of the three electronic financial statements are obtained respectively. The risk coefficients are compared with the coefficient threshold respectively, and it is obtained that electronic financial statement A is level two, electronic financial statement B is level one, and electronic financial statement C is level three.

[0080] Retrieve the audit level corresponding to the electronic financial statement, and match the corresponding audit process according to the audit level of the electronic financial statement; when the audit level of the electronic financial statement is level one, conduct a primary audit on the electronic financial statement; when the audit level of the electronic financial statement is level two, after conducting a primary audit on the electronic financial statement, send it to the financial personnel; when the audit level of the electronic financial statement is level three, send the electronic financial statement after primary audit and financial personnel audit to the final auditor, and the final auditor audits the electronic financial statement; mark the electronic financial statement passed the audit as passed, return the electronic financial statement that failed the audit, and notify the corresponding reimbursement personnel to verify and then re-upload the financial statement image.

[0081] Obtain the information of the reimbursement personnel for the electronically audited financial statements; among them, the information of the reimbursement personnel includes the department of the reimbursement personnel and the level of the reimbursement personnel; divide and number the electronically audited financial statements according to the departments of the reimbursement personnel to obtain the large groups of the electronically audited financial statements; divide the electronically audited financial statements in each large group according to the level of the reimbursement personnel to obtain the levels of the electronically audited financial statements; obtain the level encryption library; match the corresponding level encryption library according to the level of the electronically audited financial statements, and randomly select an encryption algorithm from the level encryption library to encrypt the corresponding electronically audited financial statements; construct the corresponding department blockchain according to the level of the electronically audited financial statements, and store the encrypted electronically audited financial statements in the corresponding department blockchain according to the numbers; integrate several department blockchains to obtain the financial statement blockchain.

[0082] For example: Suppose there are currently three electronically audited financial statements. According to the information of the reimbursement personnel of the electronically audited financial statements, the first and third electronically audited financial statements belong to Department 1, and the second electronically audited financial statement belongs to Department 4; then number the first and third electronically audited financial statements as A, and number the second electronically audited financial statement as D; according to the level of the reimbursement personnel, the level of the first electronically audited financial statement is A3, the level of the second electronically audited financial statement is D2, and the level of the third electronically audited financial statement is A1. The corresponding level encryption libraries are the high-level encryption library, the medium-level encryption library, and the low-level encryption library respectively; then use the corresponding low-level encryption library to encrypt the third electronically audited financial statement, use the medium-level encryption library to encrypt the second electronically audited financial statement, and use the high-level encryption library to encrypt the first electronically audited financial statement; and store the first and third electronically audited financial statements on the blockchain of the corresponding Department A, and store the second electronically audited financial statement on the blockchain of the corresponding Department D.

[0083] It should be noted that the audit level of the electronically audited financial statements is analyzed according to the risk coefficient of the electronically audited financial statements, and the corresponding audit process is matched according to the audit level; it is possible to conduct multiple audits on electronically audited financial statements with high risks, reducing errors in the audit process; for electronically audited financial statements with low risks, rapid audits are conducted, which is beneficial to improving the efficiency and quality of audits; after classifying the electronically audited financial statements that have passed the audit, encrypting and storing them in the corresponding blockchain can audit the electronically audited financial statements, which is beneficial to preventing the electronically audited financial statements from being maliciously tampered with and improving data security.

[0084] Some of the data in the above formula are taken as numerical values after removing the dimension. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0085] Working principle of the present invention: Obtain the financial statement image uploaded by the user; Generate an electronic financial statement based on the financial statement image; Analyze the conversion accuracy rate according to the electronic financial statement; Obtain the risk data of the electronic financial statement; Calculate the risk coefficient of the electronic financial statement according to the risk data; Determine the review level of the electronic financial statement based on the risk coefficient; Review the electronic financial statement according to the review level; Divide the electronically audited financial statement to obtain the level of the electronic financial statement; Encrypt and store the electronic financial statement according to the level of the electronic financial statement.

[0086] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An automatic audit management system for intelligent financial statements, characterized in that: include: Automatic audit module, and the data collection module and terminal management module connected to it; Data collection module: used to obtain financial report images uploaded by users; generate electronic financial reports based on financial report images; analyze conversion accuracy based on electronic financial reports; Automatic audit module: used to obtain risk data of electronic financial statements; Calculate the risk factor of electronic financial statements based on risk data; Determine the audit level of electronic financial statements based on the risk factor; audit the electronic financial statements according to the audit level; Terminal management module: used to classify the electronic financial statements that have passed the review to obtain the level of the electronic financial statements; and encrypt and store the electronic financial statements according to the level of the electronic financial statements.

2. According to claim 1, the automatic audit management system of intelligent financial statements is characterized in that: The step of generating an electronic financial statement based on the financial statement image includes: Retrieving financial statement images and preprocessing the financial statement images; wherein the preprocessing includes: cropping, denoising, brightness adjustment, and tilt correction; Import the financial statement image into OCR text recognition software to generate electronic financial statements; OCR text recognition software includes: Rossum or Klippa.

3. According to claim 1, the automatic audit management system of intelligent financial statements is characterized in that: The analysis of conversion accuracy based on electronic financial statements includes: Retrieve electronic financial statements and financial statement images; use functions in Python to convert electronic financial statements into images to obtain converted images; The conversion image and the financial statement image are compared to generate conversion similarity; the conversion similarity is marked as the conversion accuracy rate of the electronic financial statement.

4. The automatic audit management system for intelligent financial statements according to claim 3 is characterized in that: The method for obtaining the conversion similarity includes: Retrieving the conversion image and the financial statement image; cropping the financial statement image and the conversion image into the same size; dividing the financial statement image and the conversion image into a number of comparison areas according to the set area sizes; Matching the comparison areas of the financial statement image and the conversion image to obtain a plurality of comparison images; integrating the plurality of comparison images into a comparison input sequence; The similarity analysis model is called, and the comparison input sequence is input into the similarity analysis model to obtain the transformation similarity between the financial statement image and the transformation image; wherein the similarity analysis model is constructed based on an artificial intelligence model.

5. The automatic audit management system for intelligent financial statements according to claim 4 is characterized in that: The similarity analysis model is constructed based on an artificial intelligence model and includes: Acquire a standard data set; wherein the standard data set includes standard input data consistent with the content attributes of the comparison input sequence and standard input data consistent with the content attributes of the conversion similarity; Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence model; use the test set to test the artificial intelligence model and obtain the test indicators; Determine whether the test index is greater than the index threshold. If yes, mark the trained artificial intelligence model as a similarity analysis model; if no, retrain the artificial intelligence model; wherein the artificial intelligence model includes: a convolutional neural network model or a long short-term memory neural network model.

6. The automatic audit management system for intelligent financial statements according to claim 1 is characterized in that: The step of calculating the risk factor of the electronic financial statement based on the risk data includes: Retrieve risk data of electronic financial statements, including conversion accuracy, statement amount, and credit rating of the person reporting the financial statements; The risk factor of the electronic financial statement is calculated by mapping the conversion accuracy, statement amount, and credit rating of the account reporting personnel with the risk factor of the electronic financial statement.

7. The automatic audit management system for intelligent financial statements according to claim 6 is characterized in that: The determination of the audit level of the electronic financial statements based on the risk factor includes: Retrieving the risk coefficient of the electronic financial statement and obtaining the coefficient threshold; comparing the risk coefficient with the coefficient threshold respectively; wherein the coefficient threshold includes a primary threshold and a secondary threshold; Determine whether the risk factor is greater than the first-level threshold; if so, compare the risk factor with the second-level threshold; if not, set the audit level of the corresponding electronic financial statement to level one; Determine whether the risk factor is greater than the secondary threshold; if yes, set the audit level of the corresponding electronic financial statement to level three; if no, set the audit level of the corresponding electronic financial statement to level two.

8. The automatic audit management system for intelligent financial statements according to claim 7 is characterized in that: The audit of electronic financial statements according to the audit level includes: Retrieve the audit level corresponding to the electronic financial statements, and match the corresponding audit process according to the audit level of the electronic financial statements; When the audit level of the electronic financial statements is level one, the electronic financial statements will be subject to primary audit; when the audit level of the electronic financial statements is level two, the electronic financial statements will be subject to primary audit before being sent to the financial staff; When the audit level of the electronic financial statements is level three, the electronic financial statements that have been audited by the primary auditor and the financial staff will be sent to the final auditor, who will audit the electronic financial statements; The electronic financial statements that have passed the review will be marked as reviewed and approved, and the electronic financial statements that have not passed the review will be returned, and the corresponding accounting personnel will be notified to verify and re-upload the financial statement images.

9. The automatic audit management system for intelligent financial statements according to claim 1 is characterized in that: The electronic financial statements that have passed the audit are divided into grades of electronic financial statements, including: Obtain the account reporting personnel information of the audited electronic financial statements; the account reporting personnel information includes the account reporting personnel department and account reporting personnel grade; The electronic financial statements are divided and numbered according to the departments of the accountants to obtain large groups of the electronic financial statements; the electronic financial statements in each large group are divided according to the levels of the accountants to obtain the levels of the electronic financial statements.

10. The automatic audit management system for intelligent financial statements according to claim 9, characterized in that: The step of encrypting and storing the electronic financial statements according to the level of the electronic financial statements includes: Retrieve the level of the electronic financial statement; obtain the level encryption library; match the corresponding level encryption library according to the level of the electronic financial statement, and randomly select an encryption algorithm from the level encryption library to encrypt the corresponding electronic financial statement; Build the corresponding departmental blockchain according to the level of the electronic financial statement, store the encrypted electronic financial statement in the corresponding departmental blockchain according to the number; integrate several departmental blockchains to obtain the financial statement blockchain.

Images