Accounting financial training evaluation method and system based on neural network model, and storage medium

Abstract

The invention provides an accounting financial training evaluation method and system based on a neural network model and a storage medium, and the method comprises the steps: obtaining the login information of a training object and the last training rating information, and requesting to configure a simulation original bill to be processed of a simulation customer corresponding to the simulation customer grade; acquiring an on-site training image and a simulated original bill image; obtaining a first recognition result and a second recognition result; obtaining voucher input information manually input by the training object, and performing difference detection on the second identification result and the voucher input information in real time; when it is detected that a second behavior pre-judged by the first recognition result is finished, comparing the first-time input data after the second behavior is finished with the last-time input data before the second behavior, and if the data are not the same, sending out a first instruction; calculating a training accuracy evaluation value of the training object; a training rating for the training subject is determined. By adopting the training evaluation method provided by the invention, the training scheme can be optimized and the training quality can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of accounting training, and in particular to an accounting and financial training assessment method, system and storage medium based on a neural network model. Background Art

[0002] Financial training is the professional knowledge training for financial practitioners or those who intend to engage in financial work, including accounting, finance, taxation, auditing and other aspects, with the purpose of enabling relevant personnel to have professional financial knowledge and skills and improve financial management level. Specific training content may include accounting basics, financial management, financial statement analysis, tax planning, internal auditing and other aspects.

[0003] Among them, accountants, as an important part of financial practitioners, daily work includes preparing accounting vouchers according to the information on the original bills, registering account books, and preparing reports, etc. Therefore, various types of specialized accounting online or offline financial training have emerged in the market.

[0004] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0005] At present, the accounting and finance training on the market is still at the theoretical knowledge stage, and rarely involves practical training on the daily work of accounting. On the other hand, even after the training, an exam or assessment test can be set to test the trainees' mastery of the knowledge and skills in the training, but the content of such exams or assessment tests is often highly homogeneous, with almost no difference in difficulty, and the improvement of trainees' abilities is relatively limited.

[0006] Therefore, there is a need for an accounting and financial training and assessment method, system and storage medium based on a neural network model to at least partially solve the above technical problems. Summary of the invention

[0007] In view of this, an embodiment of the present invention provides an accounting and financial training and assessment method, system and storage medium based on a neural network model to at least solve one of the problems in the prior art.

[0008] In a first aspect, an embodiment of the present invention provides an accounting and financial training and evaluation method based on a neural network model, the training and evaluation method comprising the following steps:

[0009] The training object's login information for logging into the financial training system is obtained, the training object's most recent training rating information is obtained based on the login information, and a request is made for the training object to configure a to-be-processed simulated original bill of at least one simulated customer of a corresponding simulated customer grade determined according to a preset matching rule based on the most recent training rating information; wherein, if it is known that it is the first login of the training object, a request is made to configure a to-be-processed simulated original bill of a pre-defined first-level simulated customer;

[0010] Acquire on-site training images of the training object including bill sorting, voucher entry, bookkeeping review and simulated tax return process collected by the camera, and acquire simulated original bill images;

[0011] Performing image recognition on the on-site training image to obtain a first recognition result of the posture and process of the training subject, performing image recognition on the simulated original bill image to obtain a second recognition result of bill information including the bill issuing unit, the bill receiving unit, the invoice items and the remarks information, and classifying the simulated original bill according to a preset classification rule based on the second recognition result to obtain bill category data;

[0012] Receive a voucher generation instruction, obtain voucher entry information manually input by the trainee based on the simulated original bill, and perform a difference detection between the second recognition result and the voucher entry information in real time, so as to issue a correction instruction when each classification error or input data error is detected; detect whether the report data is updated after each voucher entry, and if not, issue a first instruction for indicating the existence of a problem;

[0013] After receiving the accounting completion instruction, the voucher input data of the simulated customer is logically checked to issue a correction instruction each time an abnormal data is found; wherein, after receiving the accounting completion instruction, if it is detected that the input data of the simulated customer is modified again, each modification of the input data is marked as the first behavior by adding a label;

[0014] Sending the voucher entry data of the simulated customer to a first audit model pre-established based on a neural network model, and outputting audit data indicating the number of errors in the voucher entry data through the first audit model;

[0015] receiving a simulated tax return instruction, sending simulated tax return data to a second audit model pre-established based on a neural network model, and outputting feedback data indicating the number of errors in the simulated tax return data through the second audit model;

[0016] When the second behavior predetermined by the first recognition result based on the on-site training image is detected to be finished, detecting whether the first input data after the second behavior is finished and the last input data before the second behavior is appeared are the same, and issuing the first instruction if they are not the same;

[0017] The number of simulated customers assigned to the training object, the number of simulated original bills, the number of vouchers generated, the number of first instructions issued, the number of correction instructions issued, the number of first actions, the number of errors in the audit data, and the number of errors in the feedback data are counted, and the training accuracy evaluation value ω of the training object is calculated according to the preset corresponding influence coefficient.

[0018] Among them, 0≤ω≤1, when the value of ω is larger, it means that the training accuracy of the training object is higher;

[0019] Compare the training accuracy evaluation value with a preset threshold value. If the training accuracy evaluation value is greater than the preset threshold value, terminate the training. Otherwise, re-request at least one simulated original bill to be processed by a simulated customer with the same simulated customer classification, and iterate the first eight steps until the training accuracy evaluation value is greater than the preset threshold value or the number of iterations reaches the preset number.

[0020] Based on the training accuracy evaluation value at the end of the training, the training rating of the trainee is determined and stored, specifically: when 0≤ω≤0.7, the training rating is simple, when 0.7<ω≤0.8, the training rating is general, when 0.8<ω≤0.9, the training rating is proficient, and when 0.9<ω≤1, the training rating is complex.

[0021] In a second aspect, an embodiment of the present invention further provides a financial training evaluation system, the training evaluation system comprising:

[0022] A memory for storing computer executable instructions;

[0023] The processor is used to implement the training and assessment method of the above technical solution when executing the computer executable instructions stored in the memory.

[0024] In a third aspect, an embodiment of the present invention further provides a storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the training and assessment method of the above technical solution.

[0025] In a fourth aspect, an embodiment of the present invention further provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the training and assessment method of the above technical solution.

[0026] According to the training assessment method of the embodiment of the present invention, the input data and posture of the trainees in the whole process of simulated original bill processing including bill sorting, voucher entry, accounting review and simulated tax reporting can be analyzed, errors can be detected and reminders can be given, and the training accuracy assessment value and training rating of the trainees that can be used to characterize the professional skills of the trainees can be obtained to determine the simulated customer classification of the next simulated original bill to be processed, so as to achieve more reasonable and healthy dynamic adjustment of training, which can more significantly improve the training effect on the trainees.

[0027] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention may be achieved and obtained by the structures specifically indicated in the specification and the accompanying drawings.

[0028] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention. The components in the drawings are not drawn to scale, but are only for illustrating the principles of the present invention. In order to facilitate the illustration and description of some parts of the present invention, the corresponding parts in the drawings may be enlarged, that is, they may become larger relative to other components in the exemplary device actually manufactured according to the present invention. In the drawings:

[0030] Figure 1 A flow chart of a training and evaluation method according to an embodiment of the present invention;

[0031] Figure 2 FIG. 4 is a schematic diagram of a training and evaluation system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0033] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0034] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.

[0035] It should also be noted that, unless otherwise specified, the term “connection” herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0036] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0037] First, refer to Figure 1 The following describes an accounting and financial training assessment method 100 based on a neural network model according to an embodiment of the present application. Figure 1 As shown, the training assessment method 100 may include the following steps:

[0038] In step S101, the login information of the trainee for logging into the financial training system is obtained. The most recent training rating information of the trainee is obtained based on the login information. Based on the most recent training rating information, a request is made to configure for the trainee at least one simulated original bill to be processed of a simulated customer of a corresponding simulated customer grade determined according to a preset matching rule. If it is known that this is the first login of the trainee, a request is made to configure a simulated original bill to be processed of a pre-defined first-level simulated customer.

[0039] In step S103, the on-site training images of the training object including bill sorting, voucher entry, accounting review and simulated tax declaration process are acquired through the camera, and the simulated original bill images are acquired.

[0040] In step S105, the on-site training image is subjected to image recognition to obtain a first recognition result regarding the posture and process of the training subject. The simulated original bill image is subjected to image recognition to obtain a second recognition result of bill information including the bill issuing unit, the bill receiving unit, the invoice items and the remarks information. Based on the second recognition result, the simulated original bill is classified according to the preset classification rules to obtain bill category data.

[0041] In step S107, a voucher generation instruction is received, and the voucher input information based on the simulated original bill manually input by the trainee is obtained. The second recognition result and the voucher input information are detected in real time to issue a correction instruction when each classification error or input data error is detected. It is detected whether the report data is updated after each voucher entry, and if not, the first instruction is issued.

[0042] In step S109, after receiving the accounting completion instruction, the voucher input data of the simulated customer is logically checked to issue a correction instruction when the data is abnormal. After receiving the accounting completion instruction, if it is detected that the input data of the simulated customer is modified again, each modification of the input data is marked as the first behavior by adding a label.

[0043] In step S111, the voucher entry data of the simulated customer is sent to a first audit model pre-established based on a neural network model, and the first audit model outputs audit data for indicating the number of errors in the voucher entry data.

[0044] In step S113, a simulated tax return instruction is received, and simulated tax return data is sent to a second audit model pre-established based on a neural network model, and feedback data for indicating the number of errors in the simulated tax return data is output by the second audit model.

[0045] In step S115, when the second behavior pre-determined based on the first recognition result of the on-site training image is detected to end, it is detected whether the first input data after the second behavior ends and the last input data before the second behavior occurs are the same. If they are not the same, a first instruction is issued.

[0046] In step S117, the number of simulated customers assigned to the training object, the number of simulated original bills, the number of generated vouchers, the number of first instructions issued, the number of correction instructions issued, the number of first actions, the number of errors in the audit data, and the number of errors in the feedback data are counted. The training accuracy evaluation value ω of the training object is calculated according to the preset corresponding influence coefficient.

[0047] Among them, 0≤ω≤1, the larger the ω value is, the higher the training accuracy of the training object is.

[0048] In step S119, the training accuracy evaluation value is compared with a preset threshold. If it is greater than the preset threshold, the training is terminated. Otherwise, a request is made to reconfigure at least one simulated original bill to be processed for the same simulated customer grade, and steps S103 to S117 are iterated until the training accuracy evaluation value is greater than the preset threshold or the number of iterations reaches the preset number.

[0049] In step S121, based on the training accuracy evaluation value at the end of the training, the training rating of the training object is determined and stored. Specifically, when 0≤ω≤0.7, the training rating is simple, when 0.7<ω≤0.8, the training rating is general, when 0.8<ω≤0.9, the training rating is proficient, and when 0.9<ω≤1, the training rating is complex.

[0050] In an embodiment of the present application, firstly, the login information of the trainee for logging into the financial training system and the most recent training rating information are obtained, and based on the most recent training rating information, a request is made to configure the simulated original bill to be processed corresponding to the simulated customer classification; the on-site training image and the simulated original bill image of the trainee captured by the camera are obtained; image recognition is performed on the on-site training image to obtain a first recognition result related to the posture of the trainee and the workflow in which the trainee is located, and image recognition is performed on the simulated original bill image to obtain a second recognition result representing the bill information; then, a difference detection is performed on the second recognition result and the voucher entry information based on the simulated original bill manually input by the trainee; after receiving the accounting completion instruction, a logical check is performed on the voucher entry data of the simulated customer; and a pre-recorded account is sent to the pre-recorded account. The first audit model established first sends the voucher entry data and receives the audit data for indicating the number of errors in the voucher entry data; then receives the simulated tax return instruction, extracts and detects the simulated customer information in any two adjacent simulated tax returns, and obtains the feedback data output by the second audit model for indicating the number of errors in the simulated tax return data; when the second behavior predetermined by the first recognition result based on the on-site training image is detected to be over, detect whether the first input data after the end of the second behavior and the last input data before the second behavior appear are the same; then calculate the training accuracy evaluation value of the training object; compare the training accuracy evaluation value of the training object with the preset threshold, perform step iteration or end the training according to the comparison result, and finally determine the training rating of the training object.

[0051] From the description of the above process, it can be seen that according to the training evaluation method 100 of the embodiment of the present application, the input data and posture of the trainee in the whole process of simulated original bill processing can be analyzed, errors can be detected and reminders can be given, and the training accuracy evaluation value and training rating of the trainee that can be used to characterize the professional skills of the trainee can be obtained to determine the simulated customer grade of the next simulated original bill to be processed. At the same time, when the training accuracy evaluation value is obtained, the training accuracy evaluation value is compared with the preset threshold. If it is greater than the preset threshold, the training is terminated, otherwise, at least one simulated original bill to be processed by the simulated customer with the same simulated customer grade is re-requested, and the first eight steps are iterated until the training accuracy evaluation value is greater than the preset threshold or the number of iterations reaches the preset number, so as to ensure that each training meets the minimum requirements; the training evaluation method 100 according to the embodiment of the present application can more significantly improve the quality of training.

[0052] Among them, Figure 1 In the figure, step S115 is shown to be arranged between step S113 and step S117, which is only an example. It can be understood that step S115 can be located at any step between step S107 and step S117.

[0053] The contents of the above steps of the training assessment method 100 according to the embodiment of the present application will be described in detail below.

[0054] First, the financial training system in the embodiment of the present application can be an existing financial training system set on a smart terminal device such as a computer, pad or mobile phone, and the trainees are trained on the simulated original bills and the financial training system at the corresponding workstations. The simulated original bills can be invoices, expense reports, payrolls or contracts, etc. The camera used to capture images in the embodiment of the present application can be a camera built into a smart device such as a computer, pad or mobile phone on which the financial training system is installed, or it can be an independent camera arranged at a workstation, which is used to capture images at the entire workstation, including user interface images on the display screen of the smart device, desktop file images, and on-site training images of accountants.

[0055] The above illustrated scenarios are merely examples and are not intended to be limiting.

[0056] In an embodiment of the present application, in step S101, the login information of the trainee for logging into the financial training system is obtained. Based on the login information, the most recent training rating information of the trainee is obtained. Based on the most recent training rating information, a request is made to configure the simulated original bills to be processed of one or more simulated customers of the corresponding simulated customer classification determined according to the preset matching rules for the trainee. If it is known that this is the first login of the trainee, a request is made to configure the simulated original bills to be processed of the pre-defined first-level simulated customers.

[0057] Specifically, the financial training system of the embodiment of the present application can be used by trainees to log in for training. The number of trainees is not limited. The financial training system can pre-assign login information to the trainees or the trainees can register login information themselves. Generally, the login information includes a login account and a login password. The login information of the trainees is stored in the financial training system.

[0058] When the trainee logs in to the financial training system on site to prepare for training, the financial training system obtains the trainee's login information. Based on the login information, the pre-stored training rating information of the trainee is retrieved. Based on the most recent training rating information, a request is made for the trainee to configure one or more simulated original bills (paper) to be processed for the simulated customer grades corresponding to the preset matching rules. Among them, the specific preset matching rules for training ratings and different simulated customer grades will be described in detail below. After the financial training system issues a request, the trainee can pick up the pre-prepared simulated original bills to be processed corresponding to the simulated customer grades by himself or other personnel, and then the trainee can prepare for the next process. Among them, the simulated customers and simulated original bills can be simulated according to actual customers and original bills.

[0059] If it is learned that this is the first time for the trainee to log into the financial training system, a request is made to configure the simulated original bills to be processed with the simulated customer graded as level one.

[0060] In the embodiment of the present application, in step S103, on-site training images of the training object including bill sorting, voucher entry, bookkeeping review and simulated tax return process are acquired through the camera, and the simulated original bill image is also acquired through the camera.

[0061] Of course, before performing step S103, it is also necessary to perform the following steps: collecting on-site training images of the training object including bill sorting, voucher entry, bookkeeping review and simulated tax return process. Collecting simulated original bill images.

[0062] The simulated original bill images can be collected by: when the trainee takes the simulated original bills to be processed and sorts the bills at the workstation, the simulated original bill images are collected one by one through a camera. There is no specific limitation.

[0063] Similarly, on-site training images can also be obtained by capturing the corresponding process on-site through cameras.

[0064] In the embodiment of the present application, in step S105, image recognition is performed on the on-site training image to obtain a first recognition result of the posture and workflow of the training object. Image recognition is performed on the simulated original bill image to obtain a second recognition result of bill information including the bill issuing unit, the bill receiving unit, the invoice items and the remarks information. Based on the second recognition result, the simulated original bill is classified according to the preset classification rules to obtain bill category data.

[0065] Specifically, the financial training system can use existing image recognition algorithms to perform image recognition on the on-site training images to obtain a first recognition result about the posture and process of the training object. The first recognition result includes the posture behavior information of the training object and the user interface display information of the financial training system. The posture behavior information of the training object is used to indicate whether the training object is present, standing, making or receiving a call, or leaving the seat. The user interface display information of the financial training system is used to indicate that the training object is in the state of entering vouchers, simulating tax returns, or other work processes.

[0066] Similarly, the financial training system can use existing image recognition algorithms to perform image recognition on simulated original bill images, and can use OCR technology to perform character recognition and information extraction to obtain a second recognition result of bill information including the issuing unit, receiving unit, invoice items and remarks information.

[0067] According to the second recognition result, the simulated original bills can be classified according to enterprise cost, enterprise expense and enterprise income, and divided into cost bills, expense bills and income bills.

[0068] In addition, image recognition can also be used to identify problematic simulated original bills and remove them. Determine the classification of simulated original bills. For simulated original bills of the invoice type, determine whether the issuer and the recipient's company names are the companies that are currently accounting in the system, whether the bill content is in line with the current company's business scope, and whether the bill date is in line with the current accounting cycle. If the simulated original bill is a payroll, determine whether the employee's name is an employee of the company, whether the number of employees is sufficient, and whether the bank card number is correct.

[0069] In the embodiment of the present application, in step S107, a voucher generation instruction is received, and voucher input information based on the simulated original bill manually input by the training subject is obtained. Difference detection is performed on the second recognition result and the voucher input information in real time, so as to issue a correction instruction when each classification error or input data error is detected. At the same time, it is detected whether the report function is turned on, and if it is turned on, the first instruction is issued.

[0070] Specifically, after the trainees have sorted out the simulated original bills to be processed at their workstations, they start to enter the voucher entry information in the financial training system by comparing them with the simulated original bills. On the one hand, the financial training system receives the voucher generation instruction and obtains the voucher entry information manually entered by the trainees. On the other hand, it simultaneously calls the second recognition result based on the image recognition of the simulated original bill, and performs real-time difference detection between the second recognition result and the manually entered voucher entry information. A correction instruction is issued each time a classification error or data entry error is detected. For example, if it is identified that the issuing unit of the invoice (simulated original bill) is the unit, then the bill category is an income bill, and it can be determined that it is an income business. However, the voucher entry information obtained by the trainees contains cost or expense items, which is a clear classification error. The financial training system judges the second recognition result and the manually entered voucher entry information word by word. When a typing error, omission or repetition is detected, it is determined to be an input data error.

[0071] In order to avoid issuing correction instructions too frequently to interfere with the normal training of the trainees, after judging that the input data is wrong, the financial training system will monitor the corresponding wrong data (error point) for a preset time when the input data is judged to be wrong, such as continuing to monitor for 6 seconds, and issue a correction instruction after detecting that the wrong data has not been corrected after 6 seconds. Alternatively, it is detected that other data has been entered based on the error point, and the correction instruction is issued after the data volume exceeds the data volume entered before the error point. Whichever of the two processing methods is triggered first will be executed.

[0072] When the voucher is entered, the system will automatically disable the report function because the report data comes from the accurate entry of the voucher. After the voucher is entered, the report data opened at this time is not updated. If the unupdated report data is used directly, serious errors will occur. Therefore, the financial training system detects whether the report data is updated after each voucher entry, and issues a first instruction if not. In an embodiment of the present application, in step S109, after receiving the accounting completion instruction, a logical check is performed on the voucher entry data of the simulated customer to issue a correction instruction when the data is found to be abnormal. Among them, if it is detected that the simulated customer's entry data is modified again after receiving the accounting completion instruction, each modification of the entry data is marked as the first behavior by adding a label.

[0073] Still refer to Figure 1 When the trainee completes the data entry of a simulated customer's voucher, the financial training system issues an accounting completion instruction. After receiving the accounting completion instruction, the financial training system performs a logical check on the simulated customer's voucher entry data to issue a correction instruction when data anomalies are found. For example, the system performs multiple logical checks, such as a comparison check between the general ledger and the detailed ledger, and a balance check of accounts, to ensure the consistency and accuracy of the data. If data anomalies including data inconsistency and data errors are detected, a correction instruction is issued.

[0074] If it is detected that the voucher entry data of a simulated customer is modified again after the accounting instruction for the simulated customer is completed, each behavior of modifying the input data is marked as the first behavior by adding a label.

[0075] In an embodiment of the present application, in step S111, the credential entry data of the simulated customer is sent to a first audit model pre-established based on a neural network model, and the first audit model outputs audit data for indicating the number of errors in the credential entry data.

[0076] Specifically, after the financial training system completes the logical verification of the simulated customer's voucher entry data, in order to ensure the accuracy of the data, it must also be audited by the first audit model based on the neural network model within the system. The first audit model can automatically audit the voucher entry data and obtain audit data that is used to point out the number of errors in the voucher entry data.

[0077] Among them, the first audit model can be obtained by modeling using existing methods. For example, some voucher entry data can be used as a training data set to continuously train the neural network model, and then some voucher entry data can be used as a test data set until the neural network model converges to meet the requirements and obtains the first audit model.

[0078] In an embodiment of the present application, in step S113, a simulated tax return instruction is received, and simulated tax return data is sent to a second audit model pre-established based on a neural network model, and feedback data for indicating the number of errors in the simulated tax return data is output by the second audit model.

[0079] Among them, after the financial training system has logically verified the input data of the simulated customer and audited the first audit model, the system generates the financial statements and files (i.e., simulated tax data) required for the simulated tax return. This content can adopt existing technology and will not be described in detail here. After the system generates the financial statements and files required for the simulated tax return, the trainee issues a simulated tax return request. The financial training system receives the simulated tax return instruction, submits the relevant financial statements and files of all simulated customers to the second audit model for review, and outputs feedback data for indicating the number of errors in the simulated tax return data through the second audit model to detect whether there are simulated tax return errors. For example, whether there are omissions or misreporting. Similarly, the second audit model can also be obtained by modeling using existing methods, which will not be described here.

[0080] In the embodiment of the present application, when the second behavior pre-determined by the first recognition result based on the on-site training image is detected in step S115, it is detected whether the first input data after the second behavior is completed is the same as the last input data before the second behavior occurs. If they are not the same, the first instruction is issued.

[0081] First, before performing step S115, the following steps need to be performed: determine whether the second behavior occurs based on the first recognition result of the on-site training image. For example, if it is detected that the trainee is in a conversation, making a phone call, or leaving the seat, or if it is detected that the user interface display information of the financial training system has not changed within the set time, it is determined that the second behavior has occurred. In this way, it is judged that the trainee encountered a sudden interruption event at work. For example, through interval image comparison, the trainee is still there in the previous image, but not in the next image, and there is only an empty chair, so it is judged that the trainee has left the seat. There was only one trainee in the previous image, and there was one more person in the next image, or the trainee picked up the phone and made a call.

[0082] When the second behavior is detected, the financial training system can automatically store the progress and input data of the current training object.

[0083] When the second behavior pre-determined by the first recognition result based on the on-site training image is detected to be finished, at the voucher entry stage, it is detected whether the first entry data after the second behavior is finished is the same as the last entry data before the second behavior occurs. If they are not the same, the first instruction is issued. That is, when the system resumes training, it automatically compares the first entry data with the last entry data before the interruption to see if they are the same, detects and prompts possible duplicate entry, and prevents duplicate ticket entry caused by sudden interruption events.

[0084] In the embodiment of the present application, in step S117, the number of simulated customers nt, the number of simulated original bills n1, the number of generated vouchers n2, the number of times a the first instruction is issued, the number of times b the correction instruction is issued, the number of times c the first action is issued, the number of errors d in the audit data, and the number of errors et in the feedback data are counted. The training accuracy evaluation value ω of the training object is calculated according to the preset corresponding influence coefficient,

[0085]

[0086] Among them, 0≤ω≤1. If ω is negative, it is still taken as 0. The larger the value of ω is, the higher the training accuracy of the training object is.

[0087] For example, a trainee handles 10 simulated customers, simulates 100 original bills, generates 60 vouchers, issues the first instruction 30 times, issues correction instructions 10 times, checks 2 errors in the data, and makes 1 simulated tax return error in one training cycle. The training accuracy assessment value of the trainee is

[0088]

[0089] =23.5%.

[0090] In an embodiment of the present application, in step S119, the training accuracy evaluation value is compared with a preset threshold. If it is greater than (i.e., greater than or equal to) the preset threshold, the training is terminated. Otherwise, a request is made to reconfigure at least one simulated original bill to be processed for the same simulated customer grade, and the first eight steps are iterated until the training accuracy evaluation value is greater than the preset threshold or the number of iterations reaches the preset number.

[0091] For example, if the training accuracy evaluation value of a certain training object (such as a trainee of a training institution, or a student of a related school) is 0.44, which is higher than the preset threshold value of 0.3 or other set thresholds, the training can be terminated. Alternatively, if the training accuracy evaluation value of the training object is 0.78, which is higher than the preset threshold value of 0.3, the training can also be terminated. If the training accuracy evaluation value of the training object is 0.25, which is lower than the preset threshold value of 0.3, and the training object is configured with a simulated original bill of a first-level simulated customer classification when logging in, then a random request is made to configure a simulated original bill of a first-level simulated customer classification, and then it is trained, and steps S103 to S117 are iterated. The training is terminated until the training accuracy evaluation value is greater than 0.3, or the number of iterations reaches 2 times or other set times. By setting it in this way, it is guaranteed that the training object can meet the minimum training requirements.

[0092] In an embodiment of the present application, in step S121, grading is performed based on the numerical value of the training accuracy evaluation value at the end of the training, and the training rating of the training object and the matching simulated customer classification are determined and stored. Specifically: when 0≤ω≤0.7, the training rating is simple, and the matching simulated customer classification is level one. When 0.7<ω≤0.8, the training rating is general, and the matching simulated customer classification is level two. When 0.8<ω≤0.9, the training rating is proficient, and the matching simulated customer classification is level three. When 0.9<ω≤1, the training rating is complex, and the matching simulated customer classification is level four.

[0093] Continuing from the above, if at the end of the training, the training subject with a training accuracy evaluation value of 23.5% is rated as simple, the next time the training subject logs into the financial training system, he will be configured to process the simulated original bill task to be processed for the simulated customer with simulated customer classification as level one.

[0094] Among them, the simulated customer classification from level one to level four indicates that the simulated customer requirements are higher. The specific classification standards are shown in the table below:

[0095]

[0096]

[0097] Among them, simulated customers with a total score of 5-8 are graded as level 1. Simulated customers with a total score of 9-12 are graded as level 2. Simulated customers with a total score of 13-16 are graded as level 3. Simulated customers with a total score of 17-20 are graded as level 4.

[0098] The following are some examples of simulated customer grading for several simulated customers:

[0099]

[0100] Based on the above description, according to the training assessment method 100 of the embodiment of the present application, the input data and posture of the trainees in the entire process of processing simulated original bills including bill sorting, voucher entry, accounting review and simulated tax reporting can be analyzed, training errors can be detected and corresponding reminders can be made, and the training accuracy assessment value and training rating of the trainees that can be used to characterize the professional skills of the trainees can be obtained to determine the simulated customer classification of the next simulated original bill to be processed, so as to achieve more reasonable and healthy dynamic adjustment of training, and can more significantly assist in improving the work skills training effect of accountants on trainees.

[0101] refer to Figure 2The training evaluation system 200 for implementing the training evaluation method according to the embodiment of the present application includes a processor 210 and a memory 220. The training evaluation system 200 may include one or more processors 210 and one or more memories 220. The memory 220 stores an executable program run by the processor 210, and when the executable program is run by the processor 210, the processor 210 executes the training evaluation method 100 according to the embodiment of the present application described above.

[0102] The processor 210 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities.

[0103] The memory 220 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 210 may run the program instructions to implement the simulated client function (implemented by the processor) in the embodiment of the present application described herein and / or other desired functions. Various applications and various data may also be stored in the computer-readable storage medium, such as various data used and / or generated by the application, etc.

[0104] The training and assessment system 200 may also include an input device and an output device, and these components are interconnected via a bus system (not shown) and / or other forms of connection mechanisms. Figure 2 The components and structures of the training evaluation system 200 shown are merely exemplary and non-limiting. The training evaluation system 200 may also have other components and structures as required.

[0105] The input device may be a device used by a user to input instructions, and may include one or more of a keyboard, a mouse, a microphone, a touch screen, etc. In addition, the input device may also be any interface for receiving information.

[0106] The output device may output various information (eg, images or sounds) to the outside (eg, a user), and may include one or more of a display, a speaker, etc. In addition, the output device may also be any other device with an output function.

[0107] Exemplarily, the example training and evaluation system 200 for implementing the training and evaluation method 100 according to the embodiment of the present application can be applied to terminal devices (such as mobile phones), tablet computers, laptops, ultra-mobile personal computers (ultra-mobile personal computers, UMPCs), handheld computers, netbooks, personal digital assistants (personal digital assistants, PDAs), wearable devices (such as smart watches, smart glasses or smart helmets, etc.), augmented reality (augmented reality, AR), simulated reality (virtual reality, VR) devices, smart home devices, car computers and other electronic devices, and the embodiments of the present application do not impose any restrictions on this.

[0108] Those skilled in the art can understand the specific training of the training and assessment system 200 for implementing the management method according to the embodiment of the present application in combination with the contents described above. For the sake of brevity, the specific details will not be repeated here, and only some main training of the processor 210 will be described.

[0109] In one embodiment of the present application, when the executable program is executed by the processor 210, the processor 210 executes the following steps:

[0110] The login information of the trainee for logging into the financial training system is obtained, and the trainee's most recent training rating information is extracted based on the login information, and based on the most recent training rating information, a request is made to configure for the trainee at least one simulated original bill to be processed of the simulated customer classification determined according to the preset matching rule. If it is learned that it is the first login of the trainee, a request is made to configure the simulated original bill to be processed of the pre-defined first-level simulated customer. The on-site training images of the trainee, including bill sorting, voucher entry, accounting review and simulated tax reporting processes, are obtained through the camera, and the simulated original bill images are obtained. Image recognition is performed on the on-site training images to obtain a first recognition result of the posture and process of the trainee, and image recognition is performed on the simulated original bill images to obtain a second recognition result of the bill information including the bill issuing unit, the bill receiving unit, the invoice items and the remarks information, and the simulated original bills are classified according to the preset classification rules based on the second recognition result to obtain bill category data. Receive voucher generation instructions, obtain voucher entry information based on simulated original bills manually input by the training object, and perform difference detection on the second recognition result and voucher entry information in real time, so as to issue a correction instruction when each classification error or input data error is detected; detect whether the report data is updated after each voucher entry, and if not, issue a first instruction for characterizing the problem. After receiving the accounting completion instruction, perform logical verification on the voucher entry data of the simulated customer, so as to issue a correction instruction when each data abnormality is found; wherein, after receiving the accounting completion instruction, it is detected that the input data of the simulated customer is modified again, and each behavior of modifying the input data is marked as the first behavior by adding a label. Send the voucher entry data of the simulated customer to the first audit model pre-established based on the neural network model, and output audit data for indicating the number of errors in the voucher entry data through the first audit model. Receive the simulated tax declaration instruction, send the simulated tax declaration data to the second audit model pre-established based on the neural network model, and output feedback data for indicating the number of errors in the simulated tax declaration data through the second audit model. When the second behavior pre-determined by the first recognition result based on the on-site training image is detected to be finished, detect whether the first input data after the second behavior is finished and the last input data before the second behavior appears to be the same, if not, issue the first instruction. Count the number of simulated customers assigned to the training object, the number of simulated original bills, the number of vouchers generated, the number of times the first instruction is issued, the number of times the correction instruction is issued, the number of first behaviors, the number of errors in the audit data, and the number of errors in the feedback data, and calculate the training accuracy evaluation value ω of the training object according to the preset corresponding influence coefficient, where 0≤ω≤1, and the larger the ω value, the higher the training accuracy of the training object.The training accuracy evaluation value is compared with the preset threshold. If it is greater than the preset threshold, the training is terminated. Otherwise, at least one simulated original bill to be processed by the simulated customer with the same simulated customer classification is requested again, and the first eight steps are iterated until the training accuracy evaluation value is greater than the preset threshold or the number of iterations reaches the preset number. Based on the training accuracy evaluation value, the training rating of the training object is determined and stored.

[0111] In addition, according to an embodiment of the present application, the present application further provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it is used to execute the corresponding steps of the training and assessment method of the embodiment of the present application. The storage medium may include, for example, a memory card of a smart phone, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disk read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.

[0112] In addition, according to an embodiment of the present application, the present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the training and assessment method of the embodiment of the present application.

[0113] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present application to this. Those of ordinary skill in the art may make various changes and modifications therein without departing from the scope and spirit of the present application. All these changes and modifications are intended to be included within the scope of the present application as required by the appended claims.

[0114] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0115] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0116] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0117] Similarly, it should be understood that in order to streamline the present application and help understand one or more of the various application aspects, in the description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, the method of the present application should not be interpreted as reflecting the following intention: the claimed application requires more features than the features clearly stated in each claim. More specifically, as reflected in the corresponding claims, the inventive point is that the corresponding technical problem can be solved with features less than all the features of a single disclosed embodiment. Therefore, the claims following the specific embodiment are hereby explicitly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present application.

[0118] It will be understood by those skilled in the art that, except for mutually exclusive features, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this specification may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature that provides the same, equivalent or similar purpose.

[0119] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present application and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

[0120] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some modules according to the embodiments of the present application. The application can also be implemented as a device program (e.g., computer program and computer program product) for executing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0121] It should be noted that the above embodiments illustrate the present application rather than limit the present application, and that those skilled in the art may design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets should not be constructed as a limitation to the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of multiple such elements. The present application may be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim that lists several devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names.

[0122] The above is only a specific implementation or description of a specific implementation of the present application, and the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. The protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. An accounting and financial training assessment method based on a neural network model, characterized in that: The training assessment methods include: The training object's login information for logging into the financial training system is obtained, the training object's most recent training rating information is extracted based on the login information, and based on the most recent training rating information, a request is made for the training object to configure at least one simulated original bill to be processed of a simulated customer of a corresponding simulated customer grade determined according to a preset matching rule; wherein, if it is known that it is the first login of the training object, a request is made to configure a simulated original bill to be processed of a pre-defined first-level simulated customer; Acquire on-site training images of the training object including bill sorting, voucher entry, bookkeeping review and simulated tax return process collected by the camera, and acquire simulated original bill images; Based on the on-site training images, the simulated original bill images and the neural network model, the following data to be evaluated are obtained: the number of bills generated by the training subject in the process including bill sorting, voucher entry, bookkeeping review and simulated tax reporting, as well as the number of first instructions issued by the financial training system in the process for characterizing the existence of problems, the number of correction instructions issued, the number of first actions marked, the number of errors in the review data for pointing out errors in the voucher entry data, and the number of errors in the feedback data for pointing out errors in the simulated tax reporting data; The number of simulated customers and the number of simulated original bills assigned to the training object are counted, and based on the data to be evaluated, the training accuracy evaluation value ω of the training object is calculated according to the corresponding preset influence coefficient; wherein 0≤ω≤1, when the value of ω is larger, it means that the training accuracy of the training object is higher; Compare the training accuracy evaluation value with the preset threshold. If it is greater than the preset threshold, end the training. Otherwise, re-request to configure other simulated original bills with the same simulated customer classification, and iterate the first three steps until the training accuracy evaluation value is greater than the preset threshold or the number of iterations reaches the preset number. Based on the training accuracy evaluation value at the end of the training, the training rating of the trainee is determined and stored. Specifically, when 0≤ω≤0.7, the training rating is simple, when 0.7<ω≤0.8, the training rating is general, when 0.8<ω≤0.9, the training rating is proficient, and when 0.9<ω≤1, the training rating is complex.

2. The training assessment method according to claim 1, characterized in that: The obtaining of the data to be evaluated based on the on-site training image, the simulated original bill image and the neural network model specifically refers to: Performing image recognition on the on-site training image to obtain a first recognition result of the posture and process of the training object, performing image recognition on the simulated original bill image to obtain a second recognition result of bill information including the bill issuing unit, the bill receiving unit, the invoice items and the remarks information, and classifying the simulated original bill according to a preset classification rule based on the second recognition result to obtain bill category data; receiving a voucher generation instruction, obtaining voucher entry information manually input by the trainee based on the simulated original bill, and performing a difference detection between the second recognition result and the voucher entry information in real time, so as to issue a correction instruction when each classification error or input data error is detected; Check whether the report data is updated after each voucher entry, and if not, issue a first instruction to indicate that there is a problem; After receiving the accounting completion instruction, the voucher input data of the simulated customer is logically checked to issue a correction instruction each time an abnormal data is found; wherein, after receiving the accounting completion instruction, if it is detected that the input data of the simulated customer is modified again, each modification of the input data is marked as the first behavior by adding a label; Sending the voucher entry data of the simulated customer to a first audit model pre-established based on a neural network model, and outputting audit data indicating the number of errors in the voucher entry data through the first audit model; receiving a simulated tax return instruction, sending simulated tax return data to a second audit model pre-established based on a neural network model, and outputting feedback data indicating the number of errors in the simulated tax return data through the second audit model; When the second behavior predetermined by the first recognition result based on the on-site training image is detected to be finished, it is detected whether the first input data after the second behavior is finished and the last input data before the second behavior occurs are the same. If they are not the same, a first instruction is issued.

3. The training assessment method according to claim 2, characterized in that: Also includes: Determining the second behavior based on the first recognition result of the on-site training image specifically includes: The first recognition result includes posture behavior information of the training subject and user interface display information of the financial training system, wherein the posture behavior information of the training subject is used to indicate that the training subject is in a state of being present, standing, talking, making or receiving a phone call, or leaving the seat, and the user interface display information of the financial training system is used to indicate that the training subject is in a state of entering vouchers, simulating tax returns, or other processes; If it is detected that the trainee is on the phone or away from the trainee, or the information displayed on the user interface of the financial training system does not change within the set time, it is determined as the second behavior; and / or When the second behavior is detected, the progress and input data of the current training object are automatically stored.

4. The training assessment method according to claim 2, characterized in that: The preset matching rule specifically refers to: When the training rating is simple, match the first level simulated customer; when the training rating is general, match the second level simulated customer; when the training rating is skilled, match the third level simulated customer; when the training rating is complex, match the fourth level simulated customer; and / or When an input data error is detected, a preset time of the corresponding error data when the input data error occurs is monitored to detect whether the error data is corrected after the preset time. If it is still not corrected, a correction instruction is issued.

5. The training assessment method according to claim 2, characterized in that: The logic check of the input data is performed to issue a correction instruction when the data is found to be abnormal, specifically referring to: Perform multiple logical checks, including comparison checks between the general ledger and detailed ledgers, and account balance checks. If data inconsistencies and errors are detected, corrective instructions will be issued.

6. The training assessment method according to claim 2, characterized in that: The classifying the simulated original bill according to the preset classification rule based on the second recognition result to obtain bill category data specifically refers to: The simulated original bills are classified according to enterprise cost, enterprise expense and enterprise income categories, and are divided into cost bills, expense bills and income bills.

7. The training assessment method according to claim 6, characterized in that: Detecting a classification error means that: the bill information in the second recognition result is logically checked with the bill category data, and if a mismatch is detected, it is a classification error.

8. The training assessment method according to claim 1, characterized in that: The training accuracy evaluation value ω of the training object is calculated using the following formula: Among them, nt is the number of simulated customers of the training object, n1 is the number of simulated original bills, n2 is the number of generated vouchers, a is the number of times the first instruction is issued, b is the number of times the correction instruction is issued, c is the number of first actions, d is the number of errors in the audit data and et is the number of errors in the feedback data.

9. A financial training assessment system, characterized in that: The training and evaluation system includes: A memory for storing computer executable instructions; The processor is used to implement the training and assessment method described in any one of claims 1 to 8 when executing the computer executable instructions stored in the memory.

10. A storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the training and assessment method according to any one of claims 1 to 8.

Images