Method and device for automatically cancelling exchange account after verification based on artificial intelligence

Abstract

The invention discloses an artificial intelligence-based method and device for automatically cancelling exchange accounts after verification. The method comprises the following steps: extracting key information characteristics of a contact account and order information characteristics of an existing order according to input contact account information and order information of the existing order; inputting the key information features of the exchange accounts into a pre-constructed classification model to obtain the types of the exchange accounts; according to the type of the exchange account, matching the key information feature of the exchange account with the order information feature of each existing order, and generating a verification matching result; according to the cancel-after-verification matching result, dividing the exchange accounts into debtor projects and creditor projects for matching and grouping to obtain matching and grouping data; and traversing all the matched grouped data, and performing dealing account dealing according to an amount equivalent dealing mode or a non-amount equivalent dealing mode to generate a dealing result report. The cancel-after-verification operation is automatically executed, so that the cancel-after-verification time and workload can be remarkably reduced, and the financial processing accuracy and efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the field of ERP-financial accounting technology, and in particular relates to a method and device for automatic write-off of current accounts receivable based on artificial intelligence. Background Art

[0002] With the expansion of enterprise scale and the increase of business volume, the demand for intelligent enterprise management software and financial management software is also increasing year by year, and the management of accounts receivable has become an important task in enterprise financial management. Account receivable write-off refers to the process of matching the paid or received funds with the account vouchers to write off the debits and credits of accounts receivable and payable generated by sales, procurement and other businesses with customers or suppliers during the production and operation process, and modifying the write-off status to facilitate the management of the status of accounts receivable. Its purpose is to ensure the accuracy and authenticity of the company's financial statements and avoid financial chaos and potential risks caused by long-term accounts receivable. Summary of the invention

[0003] In order to better realize the write-off of accounts receivable and payable, an embodiment of the present invention provides a method and device for automatic write-off of accounts receivable and payable based on artificial intelligence.

[0004] In a first aspect, an embodiment of the present invention provides a method for automatic write-off of current accounts based on artificial intelligence, comprising:

[0005] According to the input accounts receivable information and order information of existing orders, key information features of accounts receivable and order information features of existing orders are extracted;

[0006] Inputting the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable;

[0007] According to the type of accounts receivable, the key information characteristics of the accounts receivable are matched with the order information characteristics of each existing order to generate a write-off matching result;

[0008] According to the write-off matching result, the current accounts receivable are divided into debit items and credit items for matching grouping to obtain matching grouping data;

[0009] Traverse all the matching group data, write off the accounts receivable and payable according to the amount-equal write-off mode or the non-amount-equal write-off mode, and generate a write-off result report.

[0010] In one or some optional embodiments, the generating of the write-off matching result based on the type of the accounts receivable and receivable and the key information features of the accounts receivable and receivable and the order information features of each existing order includes:

[0011] For each type of accounts receivable, the feature information in the key information features of the accounts receivable and the order information of each existing order are converted into high-dimensional vector representations using the word embedding model, and the matching index between each accounts receivable and each existing order is calculated:

[0012] The order with the best matching index is matched with the accounts receivable to obtain a write-off matching result.

[0013] In one or some optional embodiments, for a current account and an existing order, the matching index of the current account and the existing order is calculated by the following formula:

[0014] in:

[0015]

[0016] In the formula, represents the bill matching index; θ is the cosine similarity; W d and W θ is the weight coefficient; G U The feature vector of the bill number representing the current bill; H U represents the order number feature vector; Represents G U and H U The covariance matrix of G V The feature vector representing the amount of the bill; H V Represents the order amount feature vector; Represents G V and H V The covariance matrix of G W H represents the aging feature vector of the current account; W represents the order aging feature vector; Represents G W and H W The covariance matrix of .

[0017] In one or some optional embodiments, the method further includes:

[0018] For accounts receivable that have not been successfully matched with any order and have the same trade unit number, as well as multiple groups of accounts receivable that match multiple orders, they are matched and grouped in the order of the accounting dates of the accounts receivable and added to the write-off matching results.

[0019] In one or some optional embodiments, traversing all the matching group data, writing off the accounts receivable and payable according to the amount-equal writing-off mode or the non-amount-equal writing-off mode, and generating a writing-off result report includes:

[0020] Traversing all the matching grouped data, in the write-off mode according to the amount equal value, performing standard write-off on the items whose unwritten-off amount on the debit side is equal to the unwritten-off amount on the credit side, and generating a write-off result report;

[0021] or,

[0022] Traverse all the matching grouped data, and in a mode of non-amount-equivalent write-off, calculate the total amount of debits to be written off or the total amount of credits to be written off, add the accounts receivable rows with smaller total amounts to be written off to a list to be written off, and traverse and accumulate the amounts of the rows with larger total amounts to be written off row by row until the accumulated amount is greater than or equal to the total amount of the voucher rows with smaller total amounts to be written off. Then, add the traversed accounts receivable voucher rows to the list to be written off, write off the list to be written off, and generate a write-off result report.

[0023] In one or some optional embodiments, the method further includes: in a non-amount-equivalent write-off mode, before writing off, determining that the write-off amount allocated to each account receivable does not exceed its outstanding amount, and that the total debit amount and the total credit amount of the write-off data of the accounts receivable are equal; if so, the corresponding accounts receivable will be written off.

[0024] In one or some optional embodiments, the method further includes: if an abnormality occurs in the matching process, a financial abnormality information report will be generated and sent externally.

[0025] In one or some optional embodiments, the method further includes:

[0026] Change the write-off indicator in the voucher of the current account with all the balances written off to the fully written-off indicator;

[0027] Change the write-off indicator in the voucher of the accounts receivable with partially written-off balance to the partial write-off indicator.

[0028] In one or some optional embodiments, the method further includes:

[0029] If a preset failure occurs during the write-off process, a database rollback of all operations will be performed.

[0030] In one or some optional embodiments, the method further includes training the classification model in the following manner:

[0031] Use the accounts receivable training data set to train the SVM classification training model and obtain the output training results;

[0032] Select the initial simplex point within the parameter range of the preset penalty parameter C and kernel parameter γ, use the simplex search algorithm, perform iterative optimization according to the output training results of the SVM model, and obtain the optimal penalty parameter C and kernel parameter γ;

[0033] Using the optimal penalty parameter C and kernel parameter γ as the initial solution for each particle in the particle swarm optimization algorithm, adopting the particle swarm optimization algorithm, iteratively updating the speed and position of the particle according to the output training result of the SVM model, and obtaining the updated penalty parameter C and kernel parameter γ;

[0034] In the iterative process, the global optimal solution of the penalty parameter C and the kernel parameter γ is continuously updated until the stopping condition is reached, and the final penalty parameter C and the kernel parameter γ are obtained;

[0035] The SVM classification training model is updated using the final penalty parameter C and the kernel parameter γ, and retrained until the SVM classification training model meets the end condition, thereby obtaining the classification model.

[0036] In one or some optional embodiments, the calculation formula of the RBF kernel function selected by the SVM classification training model is:

[0037] K(x i ·x j ) = exp(-γ||x i -x j || 2 )

[0038] The SVM classification training model and classification function are:

[0039]

[0040] Among them, γ is the kernel parameter of RBF kernel, γ>0, a i is the Lagrange coefficient, 0≤a i ≤C, i=1, 2, 3…m; C is the penalty parameter; x i and x j Each represents a vector, describing the attribute information of N-dimensional account information; i is the label, ±N represents the type of accounts receivable in the training set; b is the offset of the hyperplane; i≠j; m is a positive integer greater than 1.

[0041] In one or some optional embodiments, the method further includes testing the classification model using a test data set to obtain an evaluation test result of the classification model, and optimizing the classification model based on the evaluation test result.

[0042] In a second aspect, an embodiment of the present invention provides a device for automatically writing off current accounts based on artificial intelligence, comprising:

[0043] An extraction module, used to extract key information features of the accounts receivable and order information features of the existing orders based on the input accounts receivable information and order information of the existing orders;

[0044] A classification module, used to input the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable;

[0045] A matching module is used to match the key information characteristics of the accounts receivable with the order information characteristics of each existing order according to the type of the accounts receivable, and generate a write-off matching result;

[0046] A grouping module, used for dividing the current accounts into debit items and credit items according to the write-off matching result, and performing matching grouping to obtain matching grouping data;

[0047] The write-off module is used to traverse all the matching group data, write off the accounts receivable and payable according to the amount equivalent write-off mode or the non-amount equivalent write-off mode, and generate a write-off result report.

[0048] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for automatic write-off of accounts receivable based on artificial intelligence as described in the first aspect.

[0049] In a fourth aspect, an embodiment of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for automatic write-off of accounts receivable based on artificial intelligence as described in the first aspect is implemented.

[0050] In a fifth aspect, an embodiment of the present invention provides a computer program product comprising instructions, which, when executed on a computer device, enables the computer device to execute the method for automatic write-off of accounts receivable based on artificial intelligence as described in the first aspect.

[0051] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0052] The method for automatic write-off of accounts receivable based on artificial intelligence provided by the embodiment of the present invention extracts key information features of accounts receivable and inputs them into a pre-built classification model to obtain the type of accounts receivable, searches for orders according to the type of accounts receivable and matches them, divides the accounts receivable into debit items and credit items according to the write-off matching results, and performs matching grouping to obtain matching grouping data, traverses all matching grouping data, writes off accounts receivable according to the amount equivalent write-off mode or the non-amount equivalent write-off mode, and generates a write-off result report. The intelligent identification, classification, matching and automatic write-off of write-offable line items of accounts receivable are realized, and the method automatically performs the write-off operation, which can significantly reduce the time and workload of manual write-off, ensure the transparency and traceability of the write-off process, and improve the accuracy and efficiency of financial processing. It can handle different types of accounts receivable and receivable, adapt to different business needs, enhance the flexibility and adaptability of write-offs, reduce the workload of write-offs, improve the overall efficiency and quality of financial management, and has real-time and scalability to meet higher business needs. It brings innovative solutions to the write-off matching process in corporate financial management, which helps to improve the company's financial management level and competitiveness.

[0053] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0054] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0056] Figure 1 It is a flow chart of a method for automatic write-off of current accounts receivable based on artificial intelligence in one embodiment of the present invention;

[0057] Figure 2 The present invention is a schematic diagram of the structure of an apparatus for automatic cancellation of accounts receivable based on artificial intelligence in one embodiment of the present invention. DETAILED DESCRIPTION

[0058] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0059] The inventors have found that the traditional method of writing off accounts receivable mainly relies on manual review and matching, which is not only time-consuming and labor-intensive, but also prone to errors or omissions due to human factors. At the same time, when multiple payment and receipt vouchers are involved under the same order, the writing-off procedure faces great technical difficulties in the identification and matching process. In recent years, the application of artificial intelligence technology in the field of financial management has become increasingly widespread, providing new solutions for the writing-off of accounts receivable. Therefore, how to realize the automatic writing-off of accounts receivable based on artificial intelligence has become a technical problem that needs to be solved urgently. In view of the above problems, the present invention is proposed to provide a method and device for automatic writing-off of accounts receivable based on artificial intelligence that overcomes the above problems or at least partially solves the above problems.

[0060] Embodiment 1

[0061] In the first embodiment of the present invention, a method for automatic write-off of current accounts receivable based on artificial intelligence is provided. Figure 1 As shown, the method may include the following steps:

[0062] S101: extracting key information features of the accounts receivable and the order information features of the existing orders based on the input accounts receivable information and the order information features of the existing orders;

[0063] S102: Inputting the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable;

[0064] S103: According to the type of the accounts receivable, the key information features of the accounts receivable are matched with the order information features of each existing order to generate a write-off matching result;

[0065] S104: According to the write-off matching result, the current accounts receivable are divided into debit items and credit items for matching grouping to obtain matching grouping data;

[0066] S105: traverse all the matching group data, write off the accounts receivable and payable according to the amount-equal write-off mode or the non-amount-equal write-off mode, and generate a write-off result report.

[0067] The method for automatic write-off of accounts receivable based on artificial intelligence provided by the embodiment of the present invention extracts key information features of accounts receivable and inputs them into a pre-built classification model to obtain the type of accounts receivable, searches for orders and matches them according to the type of accounts receivable, divides the accounts receivable into debit items and credit items according to the write-off matching results, and obtains matching grouping data, traverses all matching grouping data, writes off accounts receivable according to the amount equivalent write-off mode or the non-amount equivalent write-off mode, and generates a write-off result report. Intelligent identification, classification, matching and automatic write-off of write-off items of accounts receivable are realized, which can improve the efficiency and accuracy of financial processing. It has real-time and scalability and meets higher business requirements. The method automatically performs write-off operations, which can significantly reduce the time and workload of manual write-offs, ensure the transparency and traceability of the write-off process, and improve the accuracy and efficiency of write-offs. It can handle different types of accounts receivable and adapt to different business needs, enhance the flexibility and adaptability of write-offs, reduce the workload of write-offs, improve the overall efficiency and quality of financial management, and bring innovative solutions to the write-off matching process in corporate financial management, which helps to improve the company's financial management level and competitiveness.

[0068] In the embodiment of the present invention, the accounts receivable information and order information in the above step S101 may be the original data of accounts receivable input by the user through the user interface of the enterprise resource planning (ERP) system, such as the names of the two parties to the transaction, the transaction amount, the transaction date, etc. The system may automatically call the natural language processing (NLP) technology to perform format verification on the input data to ensure that the data meets the preset format requirements. Noise in the data, such as irrelevant characters, repeated information, etc., is removed to improve the efficiency and accuracy of subsequent processing.

[0069] In the embodiment of the present invention, through the above step S101, the input accounts receivable information and the order information of the existing orders are processed one by one, and the key information features of the input accounts receivable and the key information features of each input existing order are extracted respectively. The specific implementation process may include the following steps:

[0070] (a) Analyze the input accounts receivable information using intelligent recognition, automatically identify and extract key information features of the accounts receivable information, which may include key information such as account number, bill amount and bill age, and mark, list and store the accounts receivable information one by one in a structured manner, recorded as List A, to provide a basis for subsequent matching mechanism operations.

[0071] (b) Analyze the order information of the input existing orders using intelligent recognition, automatically identify and extract the order information features of the existing orders, which may include order number, order amount, order amount and order aging information, mark, list and store the order information one by one in a structured manner, and record it as List B, to provide a basis for subsequent matching mechanism operations. Among the above order information features, the order number is the code that uniquely identifies the order and is important reference information in the automatic write-off process. The order information features may also include information such as order amount, order date, customer name, etc., which helps to verify the correlation between the order and the accounts receivable.

[0072] In an optional embodiment, the classification model in the above step S102 may be a pre-trained SVM classification model, and the types of the above accounts receivable may be accounts receivable, accounts payable, other receivables, other payables, advances received, and prepaid accounts. By inputting key information such as account number, bill amount, and bill age in the key information features of the accounts receivable information into the pre-built classification model, the types of accounts receivable can be obtained, so that the pending account data and clearing account data of the accounts receivable can be matched after classification.

[0073] In one embodiment, the above classification model can be trained in the following manner:

[0074] S201: Use the accounts receivable training data set to train the SVM classification training model and obtain output training results;

[0075] S202: selecting an initial simplex point within the parameter range of a preset penalty parameter C and a kernel parameter γ, using a simplex search algorithm, and performing iterative optimization according to the output training result of the SVM model to obtain the optimal penalty parameter C and kernel parameter γ;

[0076] S203: using the optimal penalty parameter C and kernel parameter γ as the initial solution for each particle in the particle swarm optimization algorithm, adopting the particle swarm optimization algorithm, and iteratively updating the speed and position of the particle according to the output training result of the SVM model to obtain the updated penalty parameter C and kernel parameter γ;

[0077] S204: During the iteration process, the global optimal solution of the penalty parameter C and the kernel parameter γ is continuously updated until the stop condition is reached, and the final penalty parameter C and the kernel parameter γ are obtained;

[0078] S205: Use the final penalty parameter C and the kernel parameter γ to update the SVM classification training model, and retrain until the SVM classification training model meets the end condition to obtain the classification model.

[0079] In a specific embodiment, the calculation formula of the RBF kernel function selected by the SVM classification training model is:

[0080] K(x i ·x j ) = exp(-γ||x i -x j || 2 )

[0081] The SVM classification training model and classification function are:

[0082]

[0083] Among them, γ is the kernel parameter of RBF kernel, γ>0, a i is the Lagrange coefficient, 0≤a i ≤C, i=1, 2, 3…m; C is the penalty parameter; x i and x j Each represents a vector, describing the attribute information of N-dimensional account information; i is the label, ±N represents the type of accounts receivable in the training set; b is the offset of the hyperplane; i≠j; m is a positive integer greater than 1.

[0084] In the above step S201, the SVM classification model is preliminarily trained using the training data set. At this time, the penalty parameter C and the kernel parameter γ of the SVM model may be set to default values ​​or random values. After training, the output training result of the SVM model, such as the classification accuracy, is obtained.

[0085] In the above step S202, an initial simplex point is selected within the parameter range of the preset penalty parameter C and the kernel parameter γ, which may include multiple initial points (i.e., multiple penalty parameter C and kernel parameter γ combinations), and the simplex search algorithm (Nelder-Mead algorithm) is applied to preliminarily optimize the penalty parameter C and the kernel parameter γ. In each iteration, the performance of the current simplex point is evaluated according to the output training results of the SVM model, and according to the evaluation results, the simplex is iteratively updated through operations such as reflection, expansion, indentation, and contraction, and the shape and position of the simplex are adjusted. Iterative optimization is performed until the stop condition is met (such as reaching the maximum number of iterations or the simplex volume is less than a certain threshold). The optimal penalty parameter C and kernel parameter γ found by the Nelder-Mead algorithm are recorded.

[0086] In the embodiment of the present invention, the particle swarm optimization algorithm (PSO) is further fine-tuned through the above steps S203 and S204, and the size of the particle swarm, the position and speed of the particles, the inertia weight, the learning factor and other parameters are set to initialize a group of random particles, and the optimal penalty parameter C and the kernel parameter γ are used as the initial solution of each particle in the particle swarm optimization algorithm to generate the spatial position and speed of the first generation of particle swarms; for the iterative update of the speed and position of the particles, according to the output training result of the SVM model, in each iteration, the particles will update two extreme values, one is the optimal solution of the particle itself, and the other is the optimal solution of the population as a whole, the optimal solution of the particle itself is called the individual extreme value, and the optimal solution of the population as a whole is called the global extreme value, and the particle inertia weight is optimized and adjusted. When the number of iterations is small, the inertia weight is dynamically increased, and when the number of iterations increases, the inertia weight is dynamically reduced, and at the same time, it is continuously judged whether the stop condition is met, such as reaching the maximum number of iterations, the global optimal solution is not updated within a certain number of times, etc. If it is met, the final penalty parameter C and the kernel parameter γ are obtained.

[0087] Through the above step S205: the final penalty parameter C and the kernel parameter γ are used to update the SVM classification training model, and the training data set is used again to train the updated SVM classification model. During the training process, it is constantly determined whether the model has reached the end condition, such as reaching a preset accuracy rate, the training time reaching a preset time, or achieving the optimal classification hyperplane by maximizing the classification interval, so that the sample points of different categories are separated to the maximum extent. When the end condition is met, the training is stopped to obtain the classification model.

[0088] In a specific embodiment, in order to optimize the classification model and improve the classification performance and generalization ability of the model, an evaluation may be performed using a test data set, which may specifically include the following steps:

[0089] The classification model is tested using a test data set to obtain an evaluation test result of the classification model, and the classification model is optimized according to the evaluation test result.

[0090] In the embodiment of the present invention, the parameters of the SVM classification model are optimized through the above steps in combination with the NM-PSO joint algorithm, and the optimized SVM classification model is used to classify the accounts receivable and receivable according to the account type, transaction nature and other characteristics of the accounts receivable and receivable. The classification results will be used for subsequent write-off matching. The performance of the SVM classification model is evaluated using an independent test data set, and the model is further optimized and adjusted to improve the accuracy and stability of the classification.

[0091] In an optional embodiment, in the above step S103, the generation of a write-off matching result by matching key information features of the accounts receivable with order information features of each existing order according to the type of the accounts receivable includes:

[0092] For each type of accounts receivable, the feature information in the key information features of the accounts receivable and the order information of each existing order are converted into high-dimensional vector representations using the word embedding model, and the matching index between each accounts receivable and each existing order is calculated:

[0093] The order with the best matching index is matched with the accounts receivable to obtain a write-off matching result.

[0094] In a specific embodiment, in the above step S103, for a current account and an existing order, the matching index between the current account and the existing order is calculated by the following formula:

[0095] Formula 1, where:

[0096]

[0097] In the formula, represents the bill matching index; θ is the cosine similarity; W d and W θ is the weight coefficient; G U The feature vector of the bill number representing the current bill; H U represents the order number feature vector; Represents G U and H U The covariance matrix of G V The feature vector representing the amount of the bill; H V Represents the order amount feature vector; Represents G V and H V The covariance matrix of G W H represents the aging feature vector of the current account; W represents the order aging feature vector; Represents G W and H W The covariance matrix of .

[0098] In the embodiment of the present invention, the key information features of the accounts receivable obtained in step S101 (i.e., the accounts receivable in list A) are used as the cancellation rules to perform adaptive matching of the matching set with the existing orders (i.e., the order information in list B), and intelligent matching is performed by corresponding and verifying with the order number information to generate a cancellation matching result. In order to better capture semantics and improve computing efficiency, it is convenient to match the various information features of accounts receivable with order information. For the key information features of accounts receivable and order information of orders, the word embedding model Word2Vec model is used to convert text information into a high-dimensional vector representation. These original features are converted into a high-dimensional vector representation that is more useful for the cancellation matching task, which is conducive to more accurate matching and cancellation in the future.

[0099] For each accounts receivable and an existing order, based on the above formula 1, if the calculated value of the accounts receivable matching index is lower, it means that the angle between the vectors is smaller and the directions are consistent, and the similarity between the accounts receivable and the order is higher. If the matching index value of an accounts receivable and an order among all existing orders is the lowest, then the matching index of the two is optimal, and the order with the best matching index can be matched with the accounts receivable to obtain the write-off matching result.

[0100] In a specific embodiment, it also includes:

[0101] For accounts receivable that have not been successfully matched with any order and have the same trade unit number, and for multiple groups of accounts receivable that have the same matching index as the same order, they will be matched and grouped in the order of the bookkeeping date of the accounts receivable and added to the write-off matching result. The order of the bookkeeping date of the accounts receivable voucher will be ranked first, and the earlier the aging date, the higher the priority, so that it can be written off first during the write-off process.

[0102] In an embodiment of the present invention, through the above-mentioned step S104, according to the write-off matching result, the accounts receivable are divided into debit items and credit items for matching and grouping, and the line items of the pending voucher and the line items of the receipt voucher are matched as a group to obtain matching group data. If it is determined that an abnormality occurs in the matching process, a financial abnormality information report will be generated and sent externally.

[0103] In the embodiment of the present invention, when allocating the write-off amount for each account voucher, it should be ensured that it does not exceed the account amount, and the total amount of the debit and credit sides of the account write-off data is equal. If any problem occurs during the matching process, for example, the account has not received any funds or the payment is still pending, or the account amount is incorrect, the system backend staff will receive a notification and manually check to find out the reason.

[0104] The system will generate a financial abnormality information report and notify the system backend data management personnel to perform verification work.

[0105] In an optional embodiment, in the above step S105, all matching group data are traversed, and the accounts receivable are written off according to the amount-equal write-off mode or the non-amount-equal write-off mode, and a write-off result report is generated, including:

[0106] Traverse all matching group data, and in the write-off mode based on equal amount, perform standard write-off on items where the unwritten-off amount on the debit side is equal to the unwritten-off amount on the credit side, and generate a write-off result report;

[0107] or,

[0108] Traverse all matching grouped data, and in the non-amount-equivalent write-off mode, calculate the total amount of debits to be written off or the total amount of credits to be written off, add the accounts receivable rows with smaller total amounts to be written off to the list to be written off, and traverse and accumulate the amounts of the rows with larger total amounts to be written off row by row until the accumulated amount is greater than or equal to the total amount of the voucher rows with smaller total amounts to be written off. Then, add the traversed accounts receivable voucher rows to the list to be written off, write off the list to be written off, and generate a write-off result report.

[0109] In the embodiment of the present invention, those skilled in the art can choose to write off the accounts receivable in accordance with the amount-equivalent write-off mode or the non-amount-equivalent write-off mode according to actual conditions and needs.

[0110] Among them, through the above-mentioned write-off mode based on the amount of equivalent value, the authenticity and accuracy of each transaction can be ensured, and the omission, wrong entry or duplication of accounting can be avoided, the accounts can be kept clear and transparent, and reliable data support can be provided for the financial management of the enterprise. In addition, the write-off process can be simplified. When the unwritten-off amount on the debit side is exactly equal to the unwritten-off amount on the credit side, the write-off can be carried out directly without complicated calculations and allocations, thereby improving work efficiency. At the same time, since the write-off mode based on the amount of equivalent value is relatively simple and clear, it reduces the complexity of manual operation and the possibility of error, thus helping to reduce operational risks.

[0111] Among them, the above-mentioned non-amount-equivalent write-off mode can handle the situation where the unwritten-off amounts on the debit and credit sides are not equal. By calculating the total amount to be written off and traversing and accumulating the amount line by line, unequal accounts can be flexibly handled, which helps enterprises to handle accounts receivable more accurately and reduce errors caused by mismatched amounts. According to actual conditions, the lines of accounts receivable with smaller total amounts to be written off can be included in the list of accounts to be written off first, and the optimization of funds can be achieved by traversing and accumulating the amounts of the lines with larger amounts line by line. It can also reduce the complexity of manual operations and the possibility of errors, thus helping to reduce operational risks.

[0112] In a specific embodiment, the method also includes: in a non-amount-equivalent write-off mode, before writing off, determining that the write-off amount allocated to each account receivable does not exceed its outstanding amount, and that the total debit amount and the total credit amount of the write-off data of the accounts receivable are equal; if so, the corresponding accounts receivable will be written off.

[0113] In an embodiment of the present invention, when write-off is executed, the write-off mode can be set to write-off by amount equivalent and write-off by non-amount equivalent. In the write-off mode by amount equivalent, identify the items with the same outstanding amount on the debit side as the outstanding amount on the credit side for standard write-off. If no matching amount is found, no write-off operation will be performed. If write-off by non-amount equivalent is selected, the total amount to be cleared on the debit side or the total amount to be cleared on the credit side is automatically calculated. Then, the voucher lines with a smaller total amount to be cleared are included in the list to be cleared, and the lines with a larger total amount to be cleared are traversed line by line and the amounts are accumulated. When the accumulated amount is greater than or equal to the total amount of the voucher lines with a smaller total amount to be cleared, these voucher lines are also included in the list to be cleared, and partial write-off logic is executed on the entire list to be cleared. The automatic write-off of accounts receivable based on artificial intelligence will finally generate a write-off result report, which lists in detail all the write-off information (including the written-off items, amounts, time, etc.), the information that failed to be cleared, and the error information. After manual review and confirmation, the system will automatically complete the write-off operation and generate the corresponding write-off voucher. The write-off result report is used for review to ensure the accuracy of the write-off results.

[0114] For example, taking the accounts receivable in accounts receivable as an example, when executing write-off, the total debit amount and total credit amount of the customer's accounts receivable will be calculated. First, the debit and credit amounts of the accounts receivable will be compared. If the credit amount is small, all debit line items will be traversed and added line by line until it is greater than or equal to the amount of the debit line items. If the total debit amount is 100 and the total credit amount is 80, the credit line items with smaller amounts will be included in the list to be written off. Next, the debit voucher line items will be traversed and added line by line until the amount is greater than or equal to 80. These voucher debit line items will be included in the list to be written off and the write-off procedure will be executed. If the last debit amount is 10 and the last credit amount is 5, the partial write-off logic will be executed for this voucher line item, and the clearing indicator will be changed to "partial write-off".

[0115] In a specific embodiment, the method further comprises:

[0116] Change the write-off indicator in the voucher of the current account with all the balances written off to the fully written-off indicator;

[0117] Change the write-off indicator in the voucher of the accounts receivable with partially written-off balance to the partial write-off indicator.

[0118] In an embodiment of the present invention, if the balance of an accounts receivable record is fully written off, the system will automatically change the write-off indicator under the general ledger view of the corresponding accounting document to "completely written off", and if the balance of an accounts receivable record is partially written off, the system will automatically change the write-off indicator under the general ledger view of the corresponding accounting document to "partially written off".

[0119] In a specific embodiment, the method further includes: if it is determined that a preset failure occurs during the write-off process, a database rollback operation of all operations is performed. If any failure occurs during the process, such as a program running error, the system generates an error message. Alternatively, if no matching grouping vouchers and receipt vouchers can be found during the write-off process, in order to ensure the integrity of the transactional operation, the system automatically performs a database rollback operation of all operations to ensure data security and consistency.

[0120] Embodiment 2

[0121] Based on the same inventive concept, an embodiment of the present invention also provides a device for automatic cancellation of current accounts based on artificial intelligence, referring to Figure 2 As shown, the device comprises:

[0122] The extraction module 101 is used to extract key information features of the accounts receivable and order information features of the existing orders based on the input accounts receivable information and order information of the existing orders;

[0123] A classification module 102, used to input the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable;

[0124] A matching module 103 is used to match the key information features of the accounts receivable with the order information features of each existing order according to the type of the accounts receivable, and generate a write-off matching result;

[0125] A grouping module 104 is used to divide the current accounts into debit items and credit items according to the write-off matching result, and perform matching grouping to obtain matching grouping data;

[0126] The write-off module 105 is used to traverse all the matching group data, write off the accounts receivable and payable according to the amount-equivalent write-off mode or the non-amount-equivalent write-off mode, and generate a write-off result report.

[0127] Embodiment 3

[0128] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned method of automatic write-off of accounts receivable based on artificial intelligence is implemented.

[0129] Embodiment 4

[0130] Based on the same inventive concept, a computer device is also provided in an embodiment of the present invention, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the above-mentioned method of automatic write-off of accounts receivable based on artificial intelligence is implemented.

[0131] Embodiment 5

[0132] Based on the same inventive concept, a computer program product containing instructions is also provided in an embodiment of the present invention. When the computer program product is run on a computer device, the computer device executes the method for automatic write-off of accounts receivable based on artificial intelligence as described above.

[0133] The principles of solving the problems by the above-mentioned apparatus, computer-readable storage medium, computer device and computer program product containing instructions in the embodiments of the present invention are similar to those of the above-mentioned methods, so their implementation can refer to the implementation of the above-mentioned methods, and the repeated parts will not be repeated.

[0134] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program codes.

[0135] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0136] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0138] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for automatic write-off of current accounts based on artificial intelligence, characterized in that: include: According to the input accounts receivable information and order information of existing orders, key information features of accounts receivable and order information features of existing orders are extracted; Inputting the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable; According to the type of accounts receivable, the key information characteristics of the accounts receivable are matched with the order information characteristics of each existing order to generate a write-off matching result; According to the write-off matching result, the current accounts receivable are divided into debit items and credit items for matching grouping to obtain matching grouping data; Traverse all the matching group data, write off the accounts receivable and payable according to the amount-equal write-off mode or the non-amount-equal write-off mode, and generate a write-off result report.

2. The method according to claim 1, characterized in that The method of generating a write-off matching result based on the type of the accounts receivable and receivable and the key information features of the accounts receivable and receivable and the order information features of each existing order includes: For each type of accounts receivable, the feature information in the key information features of the accounts receivable and the order information of each existing order are converted into high-dimensional vector representations using the word embedding model, and the matching index between each accounts receivable and each existing order is calculated: The order with the best matching index is matched with the accounts receivable to obtain a write-off matching result.

3. The method according to claim 2, characterized in that Also includes: For accounts receivable that have not been successfully matched with any order and have the same trade unit number, as well as multiple groups of accounts receivable that match multiple orders, they are matched and grouped in the order of the accounting dates of the accounts receivable and added to the write-off matching results.

4. The method according to claim 3, characterized in that Traverse all the matching group data, write off the accounts receivable and payable according to the amount-equivalent write-off mode or the non-amount-equivalent write-off mode, and generate a write-off result report, including: Traversing all the matching grouped data, in the write-off mode according to the amount equal value, performing standard write-off on the items whose unwritten-off amount on the debit side is equal to the unwritten-off amount on the credit side, and generating a write-off result report; or, Traverse all the matching grouped data, and in a mode of non-amount-equivalent write-off, calculate the total amount of debits to be written off or the total amount of credits to be written off, add the accounts receivable rows with smaller total amounts to be written off to a list to be written off, and traverse and accumulate the amounts of the rows with larger total amounts to be written off row by row until the accumulated amount is greater than or equal to the total amount of the voucher rows with smaller total amounts to be written off. Then, add the traversed accounts receivable voucher rows to the list to be written off, write off the list to be written off, and generate a write-off result report.

5. The method according to claim 4, characterized in that Also includes: In the non-amount-equivalent write-off mode, before writing off, it is determined that the write-off amount allocated to each account receivable does not exceed its outstanding amount, and the total debit amount and credit amount of the write-off data of the accounts receivable are equal; if so, the corresponding accounts receivable will be written off.

6. The method according to claim 1, characterized in that It also includes training the classification model in the following manner: Use the accounts receivable training data set to train the SVM classification training model and obtain the output training results; Select the initial simplex point within the parameter range of the preset penalty parameter C and kernel parameter γ, use the simplex search algorithm, perform iterative optimization according to the output training results of the SVM model, and obtain the optimal penalty parameter C and kernel parameter γ; Using the optimal penalty parameter C and kernel parameter γ as the initial solution for each particle in the particle swarm optimization algorithm, adopting the particle swarm optimization algorithm, iteratively updating the speed and position of the particle according to the output training result of the SVM model, and obtaining the updated penalty parameter C and kernel parameter γ; In the iterative process, the global optimal solution of the penalty parameter C and the kernel parameter γ is continuously updated until the stopping condition is reached, and the final penalty parameter C and the kernel parameter γ are obtained; The SVM classification training model is updated using the final penalty parameter C and the kernel parameter γ, and retrained until the SVM classification training model meets the end condition, thereby obtaining the classification model.

7. An artificial intelligence-based device for automatic cancellation of current accounts, characterized in that: include: An extraction module, used to extract key information features of the accounts receivable and order information features of the existing orders based on the input accounts receivable information and order information of the existing orders; A classification module, used to input the key information features of the accounts receivable into a pre-built classification model to obtain the type of the accounts receivable; A matching module is used to match the key information characteristics of the accounts receivable with the order information characteristics of each existing order according to the type of the accounts receivable, and generate a write-off matching result; A grouping module, used for dividing the current accounts into debit items and credit items according to the write-off matching result, and performing matching grouping to obtain matching grouping data; The write-off module is used to traverse all the matching group data, write off the accounts receivable and payable according to the amount equivalent write-off mode or the non-amount equivalent write-off mode, and generate a write-off result report.

8. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by the processor, the method for automatic cancellation of accounts receivable based on artificial intelligence as described in claims 1-6 is implemented.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by the processor, the method for automatic cancellation of accounts receivable based on artificial intelligence as described in claims 1-6 is implemented.

10. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the method for automatic cancellation of accounts receivable based on artificial intelligence as described in claims 1-6.

Images