Adaptive processing method, system and device of refund process and storage medium

Abstract

This invention relates to the field of financial process management systems, specifically providing an adaptive processing method, system, device, and storage medium for refund processes. The method includes: acquiring business characteristic data in response to the creation of a receipt registration form; determining the business type based on the business characteristic data, where the business type includes partial refunds; establishing an association between the receipt registration form and the original payment slip that has been settled; and, based on the association, only canceling the portion of the amount corresponding to the receipt amount in the original payment slip, and simultaneously updating the integrated data of the business system associated with the original payment slip. This invention achieves accurate cancellation of partial amounts on completed payment slips by intelligently identifying partial refund transactions and establishing an association with the original payment slip. Simultaneously, it employs a transaction mode to ensure data consistency across multiple systems, thereby improving the efficiency of financial refund processing.

Description

Technical Field

[0001] This invention belongs to the technical field of financial process management systems, specifically relating to an adaptive processing method, system, device, and storage medium for a refund process. Background Technology

[0002] In a company's financial management system, the payment reimbursement form is a core document recording the company's external payment transactions. It's common for companies, acting as payers, to need to refund portions of the payment to suppliers or other payees due to reasons such as quality deductions, price adjustments, or invoice errors. In existing financial management systems, once the payment reimbursement form completes the settlement process, its status is marked as "completed," and relevant data (such as payment amount, cash flow plan execution amount, budget allocation, etc.) is synchronized to multiple business systems, including the cash flow system, budget management system, and contract management system. The original document's process is finished and cannot be directly modified.

[0003] For the aforementioned partial refund business scenario, existing technology typically employs the following processing method: First, the finance personnel need to perform a full red-ink reversal operation on the completed original payment slip, that is, generate a red-ink document with the same amount as the original payment slip but in the opposite direction, to cancel all data records of the original payment slip; then, after the red-ink reversal operation is completed, the finance personnel need to fill out a new payment slip based on the actual amount payable; finally, the new slip needs to go through the complete approval process again, including the review at each level of approval node, until the final settlement is completed.

[0004] The existing technology has the following drawbacks: Once the original payment invoice has completed the settlement process, its status is marked as "completed," and the system usually does not allow direct modification of completed documents. Existing technology can only use a "full cancellation + re-reporting" approach, that is, first fully cancel the original document, and then create a new document. This approach essentially bypasses the immutability of documents, but it introduces operational redundancy and inefficiency. Summary of the Invention

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides an adaptive processing method, system, device and storage medium for refund process to solve the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides an adaptive processing method for a refund process, comprising: In response to the creation operation of the payment registration form, the business characteristic data of the payment registration form is obtained; Based on the business characteristic data, the business type of the payment registration form is determined, and the business type includes partial refund types; When the business type is the partial refund type, establish an association between the receipt registration form and the original payment form that has been settled; Based on the aforementioned relationship, only the portion of the original payment order corresponding to the amount received in the payment registration form is reversed, and the integrated data of the business system associated with the original payment order is updated simultaneously.

[0007] In one optional implementation, in response to a creation operation for a payment registration form, the business characteristic data of the payment registration form is obtained, including: Natural language processing technology is used to perform semantic recognition on the bank remarks or summary of the payment registration form, and refund keywords are extracted to obtain text features; Calculate the matching degree between the amount received in the payment registration form and the amount of historically completed payment forms to obtain the amount characteristics; The counterparty's account information in the payment registration form is compared with the preset supplier or customer master data to determine the matching degree and obtain the counterparty characteristics; Calculate the time proximity between the creation time of the payment registration form and the completion time of historically completed payment forms to obtain time-series characteristics; Identify the role of the operator who created the payment registration form to obtain behavioral characteristics.

[0008] In an optional implementation, the matching degree between the payment amount in the payment registration form and the amount of historically completed payment forms is calculated to obtain amount characteristics, including: Based on the counterparty account information in the payment registration form, historical completed payment orders are pre-screened to obtain a subset of payment orders that match the counterparty account information; Based on a preset amount index, multiple target payment orders that match the amount received are retrieved from the payment order subset; wherein, the amount index is stored in buckets or a B-tree index is built according to the payment amount. When the amount received is equal to the payment amount of any target payment order, the matching degree is determined to be an exact match; When the ratio of the received amount to the payment amount of any target payment order is within a preset ratio range, the matching degree is determined to be proportional matching.

[0009] In an optional implementation, the business type of the payment registration form is determined based on the business characteristic data, including: The business feature data is input into a pre-trained classification model; The classification model is used to perform inference calculations on the business feature data, and the probability distribution of the payment registration form belonging to each preset business type is output. The business type with the highest probability is determined as the business type of the payment registration form.

[0010] In an optional implementation, it further includes: After determining the business type, display the confidence information of the classification model for other business types; When the confidence level of the most probable service type is lower than a preset threshold, a prompt message is generated to guide the user to confirm or manually change the service type.

[0011] In an optional implementation, when the business type is the partial refund type, establishing an association between the payment registration form and the original payment form that has been settled includes: Based on the business characteristic data, match and display one or more candidate original payment orders from the original payment orders that have been settled. In response to the selection operation of the candidate original payment invoice, an association is established between the payment registration form and the selected original payment invoice; Based on the information in the original payment slip in the aforementioned relationship, verify that the receiving account and payment account in the payment registration form are consistent with the original payment path in the original payment slip.

[0012] In an optional implementation, based on the association, only the portion of the original payment order corresponding to the payment amount in the receipt registration form is reversed, and the integrated data of the business system associated with the original payment order is updated synchronously, including: A refund relationship graph is pre-constructed, which includes a refund relationship table. The refund relationship table is used to record the association between the receipt registration form and the original payment form, and to update the cumulative refunded amount and the remaining refundable amount of the original payment form in real time. Based on the aforementioned relationship, obtain the cumulative refunded amount and the remaining refundable amount of the original payment slip; The amount received in the payment registration form is added to the total amount refunded in the original payment form, and the remaining refundable amount is deducted accordingly. Query the refund relationship graph to verify whether the amount received in the payment registration form is not greater than the remaining refundable amount in the original payment form; Release the portion of the original payment slip corresponding to the amount received, and generate the corresponding reverse accounting voucher; The amount deducted from the planned fund execution amount associated with the original payment invoice corresponds to the amount received. Disconnect the original payment slip from the external business systems, including the budget management system and the contract management system. The data update operation to the external business system is executed in transaction mode. When any update operation fails, the compensation rollback of the successfully completed operation is triggered, and the status of the payment registration form is set to an abnormal pending state.

[0013] Secondly, the present invention provides an adaptive processing system for a refund process, comprising: The feature acquisition module is used to acquire the business feature data of the payment registration form in response to the creation operation of the payment registration form; The type determination module is used to determine the business type of the payment registration form based on the business characteristic data, wherein the business type includes partial refund types; The association establishment module is used to establish an association between the payment registration form and the original payment form that has been settled when the business type is the partial refund type; The process processing module is used to, based on the aforementioned relationship, only cancel the portion of the amount in the original payment order that corresponds to the amount received in the payment registration form, and simultaneously update the integrated data of the business system associated with the original payment order.

[0014] Thirdly, a device is provided, comprising: Memory for storing the adaptive processing program for the refund process; The processor, when executing the adaptive handler of the refund process, implements the steps of the adaptive processing method for the refund process as provided in the first aspect.

[0015] Fourthly, a computer-readable storage medium is provided, on which an adaptive processing program for a refund process is stored, wherein when executed by a processor, the adaptive processing program for the refund process implements the steps of the adaptive processing method for the refund process provided in the first aspect.

[0016] The beneficial effects of this invention lie in the fact that the adaptive processing method, system, device, and storage medium for refunds provided by this invention, through the association mechanism between the payment registration form and the original payment form, break the inherent pattern of the traditional "full amount reversal + re-reporting". In the prior art, processing partial refunds requires completing multiple independent steps such as full amount reversal, re-filling out forms, and re-approval, which are cumbersome and time-consuming. This invention integrates the above complex process into a one-stop operation of "one-time payment registration": the system automatically collects business characteristics and identifies partial refund types, intelligently recommends and associates with the original payment form, automatically inherits account information and completes amount verification, and finally achieves accurate reversal of partial amounts and synchronization of data across multiple systems. Throughout the process, users do not need to repeatedly fill out reporting forms, go through a complete approval process, or switch between multiple systems. The manual intervention steps are reduced from more than ten steps in the traditional solution to three to five steps, and the approval nodes are compressed from multiple levels to one or two levels, improving processing efficiency by several times. This fundamentally solves the technical problems of cumbersome operation and low efficiency in the prior art. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic flowchart of a method according to an embodiment of the present invention.

[0019] Figure 2 This is a schematic flowchart illustrating the process of confirming the business type according to an embodiment of the present invention.

[0020] Figure 3 This is a schematic flowchart illustrating the reversal process of a method according to an embodiment of the present invention.

[0021] Figure 4 This is a schematic block diagram of a system according to an embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of the structure of a device provided in an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0025] The adaptive processing method for the refund process provided in this embodiment of the invention is executed by a computer device, and correspondingly, the adaptive processing system for the refund process runs in the computer device.

[0026] Figure 1 This is a schematic flowchart illustrating a method according to an embodiment of the present invention. Wherein, Figure 1 The executing entity can be an adaptive processing system for a refund process. Depending on different needs, the order of steps in this flowchart can be changed, and some steps can be omitted.

[0027] like Figure 1 As shown, the method includes: S1. In response to the creation operation of the payment registration form, obtain the business characteristic data of the payment registration form; S2. Based on the business characteristic data, determine the business type of the payment registration form, wherein the business type includes partial refund types; S3. When the business type is the partial refund type, establish the association between the receipt registration form and the original payment form that has been settled; S4. Based on the aforementioned relationship, only the portion of the original payment order corresponding to the amount received in the payment registration form is cancelled, and the integrated data of the business system associated with the original payment order is updated synchronously.

[0028] In one embodiment of the present invention, based on steps S1 and S2, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner, such as... Figure 2 As shown.

[0029] In one specific implementation, when a user creates a payment registration form, the system collects multi-dimensional business characteristic data in real time to provide input for subsequent business type identification. This includes the following steps: Step S101: Obtain text features The system uses natural language processing technology to perform semantic recognition on bank remarks or summaries of payment registration forms. Specifically, the system has a pre-built refund keyword library, which includes words such as "refund," "return," "deduction," "quality deduction," and "price adjustment." The system performs word segmentation on the bank remarks and calculates the weight score of the remarks containing refund keywords.

[0030] in, Indicates the first One refund keyword, The preset weight for this keyword. This is an indicator function that takes a value of 1 when the keyword appears in the comment, and 0 otherwise. The final result is... As a text feature value, a higher score indicates that the payment registration form is more likely to be a refund transaction.

[0031] Step S102: Obtain monetary features The system calculates the match between the payment amount on the payment registration form and the amount on historically completed payment orders. To improve matching efficiency, the system uses a combination of pre-screening and index retrieval: First, based on the counterparty account information in the payment registration form, historical completed payment orders are pre-screened to obtain a subset of payment orders that match the counterparty account information. This pre-screening operation is achieved through precise or fuzzy matching of key fields such as account name and account number, narrowing the comparison scope from all historical documents to documents related to the current counterparty.

[0032] Secondly, based on a preset amount index, multiple target payment orders matching the received amount are retrieved from the payment order subset. The amount index is stored in buckets according to the payment amount or a B-tree index is built to achieve fast retrieval. For example, when using bucket storage, the payment amount is divided into intervals, such as [0,1000), [1000,5000), [5000,10000), [10000,∞), etc. During retrieval, the bucket corresponding to the interval of the received amount is directly located, and only the payment orders within that bucket are compared in terms of amount.

[0033] For the retrieved target payment order, the system calculates the amount matching degree. : When the amount received Payment amount with the target payment order When they are equal, the match is determined to be an exact match, and the match degree value is 1.

[0034] When the ratio of the received amount to the payment amount of the target payment order falls within a preset ratio range, the matching degree is determined to be proportional matching. The preset ratio range includes, but is not limited to, common refund ratios such as 50% and 80%. The matching degree calculation formula is as follows:

[0035] in, The formula represents the degree of closeness between the received amount and the expected percentage (e.g., 0.5, 0.8), where the percentage is a preset ratio. The smaller the deviation, the higher the matching degree. When there is an exact match, the matching degree is 1; when the percentage matches and fully conforms to the preset ratio, the matching degree is also 1; when the received amount deviates significantly from the preset percentage, the matching degree decreases accordingly.

[0036] The system sorts the target payment orders according to their matching degree and selects one or more historical completed payment orders with the highest matching degree as candidate associated documents for the user to choose from.

[0037] Step S103: Obtain counterparty characteristics The system compares the recipient's account information in the payment registration form with pre-set supplier or customer master data to determine the match rate. Specifically, the recipient's account information includes fields such as account name, account number, and unified social credit code. The system calculates the match rate with the master data. :

[0038] in, The similarity of account names can be calculated using edit distance or cosine similarity. The exact match score for the account (1 for a perfect match, 0 otherwise). To standardize the matching accuracy of social credit codes, , , Here are the weight coefficients for each field, and A higher match rate indicates that the counterparty of the payment record is more consistent with the known suppliers or customers in the historical master data.

[0039] Step S104: Obtain temporal features The system calculates the time proximity between the creation time of a payment receipt and the completion time of a historically completed payment receipt. Let the creation time of the payment receipt be... The completion time of the historical completed payment order is Then the time difference Time proximity The calculation formula is:

[0040] in, The attenuation coefficient is... The calculation is in days. This formula indicates that the closer the creation time of a payment receipt is to the completion time of the original payment invoice, the greater the likelihood that the payment receipt is a refund transaction. For example, when... hour, ;when Timing, The value varies It increases and then decays exponentially.

[0041] Step S105: Obtain behavioral characteristics The system identifies the role of the operator who creates the payment registration form to obtain behavioral characteristics. Specifically, the system pre-defines the mapping relationship between operator roles and behavioral weights, such as the weight corresponding to a finance specialist. Weighting of ordinary salespersons System administrator corresponding weight Behavioral characteristic values The preset weight corresponding to the operator role is directly taken, indicating that the probability of the business type being refund differs when operators with different roles create a payment registration form.

[0042] Through the above steps S101 to S105, the system obtains multi-dimensional business feature data including text features, monetary features, counterparty features, time series features, and behavioral features, forming a feature vector. This serves as input to the subsequent classification model, used to determine the business type of the payment registration form.

[0043] In one specific implementation, the system uses a pre-trained classification model to perform inference calculations on the collected multi-dimensional business feature data to automatically identify the business type of the payment registration form. Specifically, this includes the following steps: Step S201: Input the business feature data into the pre-trained classification model. The system will assemble the business feature data obtained from steps S101 to S105 into feature vectors. This serves as the input to the classification model. The classification model is a pre-trained lightweight classification model, including but not limited to logistic regression models, XGBoost models, or small neural network models.

[0044] Taking the logistic regression model as an example, the model expression is:

[0045] in, The input feature vector, The total number of preset business types, and The first The weight vector and bias terms corresponding to the business type. Representing the eigenvector Belongs to the The probability of a business type.

[0046] Taking the XGBoost model as an example, the model expression is:

[0047] in, For the number of decision trees, For the first The prediction function of a decision tree. For the first The predicted output vector for each sample, with each dimension corresponding to the predicted score for different business types.

[0048] During the model training phase, the system uses historically labeled payment registration data as training samples. These training samples include business feature data and their corresponding real business type labels (such as normal payment, full refund, partial refund, etc.). The model parameters are optimized by minimizing the loss function until the model converges.

[0049] Step S202: Perform inference calculations on the business feature data using the classification model, and output the probability distribution of the payment registration form belonging to each preset business type. The classification model takes the input feature vector Perform forward propagation calculations to output the probability distribution of the payment registration form belonging to each preset business type. The preset business types include at least normal payment type, full refund type, and partial refund type.

[0050] Let the preset business type set be The model output probability distribution is:

[0051] in, This indicates that the payment registration form belongs to the [number missing] category. The probability of the business type, and satisfying the following conditions. .

[0052] For example, for a payment receipt, the model might output the following probability distribution: normal payment probability 0.15, full refund probability 0.20, and partial refund probability 0.65. This distribution indicates that the model judges the payment receipt as a partial refund type with a high confidence level.

[0053] Step S203: Determine the business type with the highest probability as the business type of the payment registration form. Based on the probability distribution output in step S202, the system selects the business type with the highest probability value as the final business type for the payment registration form.

[0054] Using the example above, since the probability of a partial refund is the highest at 0.65, the system will determine the business type of this payment registration form as "partial refund".

[0055] In a preferred embodiment, after determining the business type, the system also displays the confidence level information of the classification model for other business types. When the confidence level of the business type with the highest probability falls below a preset threshold (e.g., 0.6), the system generates a prompt message to guide the user to confirm or manually change the business type. For example, if the model outputs a partial refund probability of only 0.55, the system may prompt: "The system speculates that this is a partial refund. Do you want to link it to the payment order? You can also manually change the business type." This mechanism makes the decision-making of artificial intelligence transparent and operable, improving the reliability of the system and the user experience.

[0056] In another preferred embodiment, the system also supports incremental learning and dynamic updates of the classification model. When a user manually corrects the model's prediction results, the system uses the corrected samples as new training data to perform incremental training on the model periodically or in real time, enabling the model to adapt to changes in user operating habits and business rules, and continuously improve classification accuracy.

[0057] In one embodiment of the present invention, based on step S3, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0058] Step S301: Based on the business characteristic data, match and display one or more candidate original payment slips from the original payment slips that have been settled. Once the system determines that the business type of the payment registration form is a partial refund, the system intelligently matches candidate documents from the original payment forms that have been settled, based on the business characteristic data obtained in steps S101 to S105.

[0059] Specifically, the system first pre-screens historical completed payment orders based on the counterparty account information in the payment registration form. The system extracts the payer's account name and number from the payment registration form and compares it with the payee account information in historical completed payment orders, filtering out a subset of payment orders with the same counterparty. This pre-screening operation significantly narrows the matching scope from all historical documents to documents related to the current counterparty, substantially improving subsequent matching efficiency.

[0060] In the pre-screened payment order set, the system calculates the matching degree by combining monetary and temporal characteristics. Let the payment amount of the payment registration order be... The payment amount of a certain original payment slip is The payment completion time is The collection registration form was created at the time of... Then the overall matching degree The calculation formula is:

[0061] in, The amount matching function is defined as follows:

[0062] in, For a preset proportion set, including but not limited to Common refund percentages, The preset tolerance threshold; and Let be the weighting coefficient, satisfying ; This is the time decay coefficient.

[0063] The system calculates the overall matching degree for each original payment slip in the payment slip subset. The system sorts the original payment slips according to their matching degree from highest to lowest. The top N original payment slips (N is a preset positive integer, such as 3 slips) with the highest matching degree are displayed as candidate documents in the user interface as a list or card.

[0064] In a preferred embodiment, the candidate document displays at least the following information: original payment order number, payment amount, payee name, payee account number, payment time, total refunded amount, and remaining refundable amount. The system also visually displays the percentage of the remaining refundable amount relative to the original payment amount using a progress bar or percentage display, helping users quickly determine whether the refund can be accepted by the current payment order.

[0065] Step S302: In response to the selection operation of the candidate original payment invoice, establish the association between the payment registration form and the selected original payment invoice. After a user selects the original payment order to be associated from the candidate document list, the system responds to the selection operation by establishing an association between the payment registration form and the selected original payment order.

[0066] Specifically, the system creates a related record in the background database, which is stored in the refund relationship table. This related record contains at least the following fields: Payment Registration Form ID, Original Payment Form ID, Refund Amount, Related Time, Related Operator, and Related Status. Through this related record, the system establishes a reverse logical link between the new payment registration form and the old payment form.

[0067] While establishing the relationship, the system updates the refund relationship graph in real time. The system can query the current cumulative refund amount for this original payment order. and remaining refundable amount The calculation formula is as follows:

[0068] in, This is the total payment amount from the original payment slip. The system will verify the refund amount. Does it meet the requirements? If the conditions are not met, an error message will be generated and the association will be prevented from being established.

[0069] In a preferred embodiment, when the original payment order selected by the user has multiple refund records, the system displays the historical refund records in the form of a timeline on the candidate display interface, including the amount, time, and corresponding payment registration number of each refund, so that the user can clearly understand the refund history of the payment order and avoid duplicate refunds or excessive refunds.

[0070] Step S303: Based on the information of the original payment slip in the aforementioned relationship, verify that the receiving account and payment account in the payment registration form are consistent with the original payment path in the original payment slip. To ensure that funds are returned to their original source and to prevent funds from flowing to the wrong destination, the system performs strict account mirroring verification on the account information in the payment registration form after establishing the association.

[0071] Let the payment account of the original payment slip be... The receiving account is The receiving account on the payment registration form is... The payment account is The system executes the following verification logic: The validation passes if both of the following conditions are met:

[0072]

[0073] That is, the receiving account on the receipt registration form must be equal to the payment account on the original payment slip to ensure that the funds are returned to the company's original payment account; at the same time, the payment account on the receipt registration form must be equal to the receiving account on the original payment slip to ensure that the refunding party is consistent with the original receiving party.

[0074] During the verification process, if any condition is not met, the system immediately generates an error message and prevents the user from continuing. The error message is described in readable natural language, such as "Verification failed: The payee account on the payment registration form should match the payment account on the original payment form" or "Verification failed: The payment account on the payment registration form should match the payee account on the original payment form," and the mismatched fields are highlighted to guide the user to correct them.

[0075] In a preferred embodiment, to strengthen the "refund to original payment method" constraint and simplify user operations, the system automatically sets the receiving account and payment account in the payment registration form to a read-only locked state after verification, prohibiting users from modifying these two fields in subsequent operations. If a user needs to modify the account information, they must first unlink it from the current original payment order, select another original payment order, or change the business type, thereby ensuring that every partial refund operation strictly follows the fund security rules of refund to the original payment method.

[0076] In one embodiment of the present invention, based on step S4, the following will provide a possible embodiment and its specific implementation will be described in a non-limiting manner, such as... Figure 3 As shown.

[0077] Step S401: Pre-build a refund relationship graph A pre-constructed refund relationship graph is used to record the association between each partial refund operation and the original payment order, and to track the refund status of the original payment order in real time. The refund relationship graph includes a refund relationship table, which contains at least the following fields: associated record ID, payment registration form ID, original payment order ID, refund amount, refund time, and association status.

[0078] The core function of the refund relationship graph is to maintain the cumulative refunded amount and remaining refundable amount for each original payment invoice. For any original payment invoice, the system maintains the following two key indicators in real time:

[0079]

[0080] in, This is the total payment amount on the original payment slip. This is the sum of refund amounts from all payment records associated with this payment order. This represents the remaining amount that can still be refunded. This graph inherently supports multiple, separate partial refunds for the same original payment invoice.

[0081] Step S402: Based on the aforementioned relationship, obtain the cumulative refunded amount and remaining refundable amount of the original payment slip. Once the payment registration form is approved, the system, based on the relationship established in step S302, queries and retrieves the current cumulative refund amount of the associated original payment invoices from the refund relationship graph. and remaining refundable amount .

[0082] Step S403: Add the amount received on the payment registration form to the cumulative refunded amount on the original payment form, and deduct the remaining refundable amount accordingly. The system calculates the updated total refunded amount and remaining refundable amount in memory:

[0083]

[0084] in, This is the amount received in the payment registration form. This update operation is calculated only in memory before the transaction is committed, and is persisted to the refund relationship graph after subsequent verification.

[0085] Step S404: Query the refund relationship graph and verify whether the amount received on the payment registration form is not greater than the remaining refundable amount on the original payment form. Before performing the actual data update operation, the system performs a validity check on the amount:

[0086] If the condition is not met, i.e. the refund amount is greater than the remaining refundable amount of the original payment order, the system will generate a clear error message, such as "Error: The refund amount (¥XXX) exceeds the remaining refundable amount (¥YYY) of the original payment order (No. PO001). Please adjust the refund amount or select another payment order", and terminate the operation.

[0087] If the verification passes, the system will continue to perform subsequent reversal and synchronization operations.

[0088] Step S405: Release the portion of the original payment slip corresponding to the amount received, and generate the corresponding reverse accounting voucher. The system releases the portion of the original payment slip corresponding to the amount received. Specifically, the system deducts the amount already held in the original payment slip. This generates a corresponding reverse accounting voucher. This reverse accounting voucher has the opposite debit and credit direction to the original payment voucher, and the amount is... It is used to accurately reflect partial refund transactions in the financial accounting system.

[0089] Step S406: Deduct the amount of the planned funding associated with the original payment invoice. The system retrieves the fund plan execution record associated with the original payment invoice and deducts the fund plan execution amount. Let the original planned funding amount be... The updated execution amount of the funding plan is as follows:

[0090] This operation ensures that the funding plan system can accurately reflect actual fund expenditures and avoid inconsistencies between funding plan execution data and actual financial data.

[0091] Step S407: Decouple the original payment slip from the external business system. The system disconnects the original payment slip from external business systems, including a budget management system and a contract management system. Specifically: For budget management systems, the system releases the budget amount occupied by the original payment invoices. The corresponding portion will be made available again.

[0092] For the contract management system, the system updates the contract execution status and deducts the accumulated payment amount from the contract. The system will then unbind this portion of the payment amount from the contract line item. If the contract has payment frequency or payment ratio restrictions, the system will synchronously update the relevant counters and ratio values.

[0093] Step S408: Execute the data update operation on the external business system in transaction mode. If any update operation fails, trigger a compensation rollback for the successful operation and set the status of the payment registration form to an abnormal pending state. The system employs the Saga transaction model to ensure eventual consistency of cross-system data updates. Specifically, the system breaks down the cross-system update operations in steps S405 to S407 into multiple compensable sub-transactions, each of which defines a corresponding compensation operation.

[0094] Let the set of sub-transactions be The corresponding compensation operation is The system executes each sub-transaction sequentially: implement (Release the amount occupied, generate reverse vouchers). If successful, continue; if it fails, trigger [the following]. (If already executed) and terminate the process; implement (Deduct the amount of funds to be executed according to the plan). If successful, continue; if unsuccessful, trigger the following steps in sequence. , Perform a compensation rollback; implement (Disconnect from the budget system). If successful, continue; if unsuccessful, trigger sequentially. , , ; implement (Disconnect from the contract management system). If successful, submit all operations; if unsuccessful, trigger sequentially. , , , .

[0095] Compensation operation Specific content and corresponding sub-transactions The operation is the opposite, for example: : Cancel the reverse accounting entry and restore the amount used; : The amount of funds to be restored according to the plan; : Restore budget allocation; : Restore the contract payment status.

[0096] When any sub-transaction fails and triggers a compensation rollback, the system sets the status of the payment registration form to "abnormal pending," records the reason for the failure and the steps taken, and notifies the system administrator for manual intervention. Simultaneously, the system retains records of the executed compensation operations to facilitate troubleshooting and data recovery.

[0097] After all sub-transactions have been successfully executed, the system will calculate the value in step S403. and Persist to the refund relationship graph, complete the entire partial refund process, and update the status of the payment registration form to "completed".

[0098] In some embodiments, the adaptive processing system for the refund process may include multiple functional modules composed of computer program segments. The computer programs for each program segment in the adaptive processing system for the refund process may be stored in the memory of a computer device and executed by at least one processor to perform (see details). Figure 1 (Description) The function of adaptive processing of the refund process.

[0099] In this embodiment, the adaptive processing system for the refund process can be divided into multiple functional modules based on the functions it performs, such as... Figure 4 As shown. The module referred to in this invention is a series of computer program segments that can be executed by at least one processor and perform a fixed function, and is stored in memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0100] The feature acquisition module is used to acquire the business feature data of the payment registration form in response to the creation operation of the payment registration form; The type determination module is used to determine the business type of the payment registration form based on the business characteristic data, wherein the business type includes partial refund types; The association establishment module is used to establish an association between the payment registration form and the original payment form that has been settled when the business type is the partial refund type; The process processing module is used to, based on the aforementioned relationship, only cancel the portion of the amount in the original payment order that corresponds to the amount received in the payment registration form, and simultaneously update the integrated data of the business system associated with the original payment order.

[0101] Figure 5 The adaptive processing method for the refund process provided in this application embodiment can be applied to a device. Those skilled in the art will understand that the device structure involved in the embodiments of this invention does not constitute a limitation on the device. The device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. Specifically, the device 500 may include: a processor 510, a memory 520, and a communication unit 530. These components communicate via one or more buses. Those skilled in the art will understand that the server structure shown in the figures does not constitute a limitation on the invention; it may be a bus topology or a star topology, and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0102] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, which, when executed, may include some or all of the steps provided in the embodiments of the present invention. The storage medium may be a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.

[0103] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0104] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or modules may be electrical, mechanical, or other forms.

[0105] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0106] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0107] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the present invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the present invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should also be covered within the protection scope of the present invention.

Claims

1. An adaptive processing method for a refund process, characterized in that, include: In response to the creation operation of the payment registration form, the business characteristic data of the payment registration form is obtained; Based on the business characteristic data, the business type of the payment registration form is determined, and the business type includes partial refund types; When the business type is the partial refund type, establish an association between the receipt registration form and the original payment form that has been settled; Based on the aforementioned relationship, only the portion of the original payment order corresponding to the amount received in the payment registration form is reversed, and the integrated data of the business system associated with the original payment order is updated simultaneously.

2. The method according to claim 1, characterized in that, In response to the creation operation of a payment registration form, the business characteristic data of the payment registration form is obtained, including: Natural language processing technology is used to perform semantic recognition on the bank remarks or summary of the payment registration form, and refund keywords are extracted to obtain text features; Calculate the matching degree between the amount received in the payment registration form and the amount of historically completed payment forms to obtain the amount characteristics; The counterparty's account information in the payment registration form is compared with the preset supplier or customer master data to determine the matching degree and obtain the counterparty characteristics; Calculate the time proximity between the creation time of the payment registration form and the completion time of historically completed payment forms to obtain time-series characteristics; Identify the role of the operator who created the payment registration form to obtain behavioral characteristics.

3. The method according to claim 2, characterized in that, Calculate the matching degree between the payment amount in the payment registration form and the amount of historically completed payment orders to obtain amount characteristics, including: Based on the counterparty account information in the payment registration form, historical completed payment orders are pre-screened to obtain a subset of payment orders that match the counterparty account information; Based on a preset amount index, multiple target payment orders that match the amount received are retrieved from the payment order subset; wherein, the amount index is stored in buckets or a B-tree index is built according to the payment amount. When the amount received is equal to the payment amount of any target payment order, the matching degree is determined to be an exact match; When the ratio of the received amount to the payment amount of any target payment order is within a preset ratio range, the matching degree is determined to be proportional matching.

4. The method according to claim 1, characterized in that, Based on the aforementioned business characteristic data, the business type of the payment registration form is determined, including: The business feature data is input into a pre-trained classification model; The classification model is used to perform inference calculations on the business feature data, and the probability distribution of the payment registration form belonging to each preset business type is output. The business type with the highest probability is determined as the business type of the payment registration form.

5. The method according to claim 4, characterized in that, The method further includes: After determining the business type, display the confidence information of the classification model for other business types; When the confidence level of the most probable service type is lower than a preset threshold, a prompt message is generated to guide the user to confirm or manually change the service type.

6. The method according to claim 1, characterized in that, When the business type is the partial refund type, establish the association between the receipt registration form and the original payment form that has been settled, including: Based on the business characteristic data, match and display one or more candidate original payment orders from the original payment orders that have been settled. In response to the selection operation of the candidate original payment invoice, an association is established between the payment registration form and the selected original payment invoice; Based on the information in the original payment slip in the aforementioned relationship, verify that the receiving account and payment account in the payment registration form are consistent with the original payment path in the original payment slip.

7. The method according to claim 1, characterized in that, Based on the aforementioned relationship, only the portion of the original payment order corresponding to the amount received in the payment registration form is reversed, and the integrated data of the business system associated with the original payment order is updated synchronously, including: A refund relationship graph is pre-constructed, which includes a refund relationship table. The refund relationship table is used to record the association between the receipt registration form and the original payment form, and to update the cumulative refunded amount and the remaining refundable amount of the original payment form in real time. Based on the aforementioned relationship, obtain the cumulative refunded amount and the remaining refundable amount of the original payment slip; The amount received in the payment registration form is added to the total amount refunded in the original payment form, and the remaining refundable amount is deducted accordingly. Query the refund relationship graph to verify whether the amount received in the payment registration form is not greater than the remaining refundable amount in the original payment form; Release the portion of the original payment slip corresponding to the amount received, and generate the corresponding reverse accounting voucher; The amount deducted from the planned fund execution amount associated with the original payment invoice corresponds to the amount received. Disconnect the original payment slip from the external business systems, including the budget management system and the contract management system. The data update operation to the external business system is executed in transaction mode. When any update operation fails, the compensation rollback of the successfully completed operation is triggered, and the status of the payment registration form is set to an abnormal pending state.

8. An adaptive processing system for a refund process, characterized in that, include: The feature acquisition module is used to acquire the business feature data of the payment registration form in response to the creation operation of the payment registration form; The type determination module is used to determine the business type of the payment registration form based on the business characteristic data, wherein the business type includes partial refund types; The association establishment module is used to establish an association between the payment registration form and the original payment form that has been settled when the business type is the partial refund type; The process processing module is used to, based on the aforementioned relationship, only cancel the portion of the amount in the original payment order that corresponds to the amount received in the payment registration form, and simultaneously update the integrated data of the business system associated with the original payment order.

9. An adaptive processing device for a refund process, characterized in that, include: Memory for storing the adaptive processing program for the refund process; A processor, when executing an adaptive handler for the refund process, implements the steps of the adaptive processing method for the refund process as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, The readable storage medium stores an adaptive processing program for the refund process, which, when executed by a processor, implements the steps of the adaptive processing method for the refund process as described in any one of claims 1-7.

Images