Method and system for public cost allocation processing

Abstract

This invention provides a method and system for public cost allocation, relating to the field of cost allocation technology. The method includes: determining a target allocation scenario based on the characteristics of the public cost invoice to be allocated; calculating an initial cost allocation ratio by calling the corresponding allocation strategy model according to the target allocation scenario and relevant original data, and simultaneously constructing an evidence chain containing snapshots of the original data; performing conflict resolution processing based on the initial ratio and the real-time budget status of each responsible entity to generate an executable cost allocation ratio; and finally completing the allocation based on the executable ratio and the evidence chain. This invention solves the technical problems of low efficiency and large errors in existing cost allocation processes, achieving automation, interpretability, and traceability of the cost allocation process, significantly improving the efficiency and accuracy of cost allocation.

Description

Technical Field

[0001] This invention relates to the field of cost-sharing technology, and more specifically, to a method and system for sharing public costs. Background Technology

[0002] In modern business activities, cost allocation is a crucial component of corporate financial management. This is especially true in group companies or shared service center models, where cost allocation among different departments, projects, or business units directly impacts the accuracy of cost accounting and the fairness of performance evaluation. Traditional cost allocation methods typically employ fixed-ratio allocation, per-capita average allocation, or allocation based on simple business volume indicators. While these methods are simple to implement, they struggle to adapt to dynamically changing business needs, and the allocation results often fail to accurately reflect the true business scenario. For example, when different business lines exhibit significant differences in the actual frequency of IT resource usage, using a uniform ratio can distort costs. Furthermore, when a shared service department supports multiple projects, manually calculating the actual service hours for each project is not only inefficient but also prone to subjective errors. In other words, existing technologies for allocating shared costs suffer from inefficiencies and significant errors. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for public cost allocation, so as to alleviate the technical problems of low allocation efficiency and large error in the prior art.

[0004] In a first aspect, embodiments of the present invention provide a method for public cost allocation, the method comprising: determining a target allocation scenario based on the characteristics of the public cost invoice to be allocated; calling a corresponding allocation strategy model according to the target allocation scenario and related original data to calculate the initial cost allocation ratio among the responsible entities, and simultaneously constructing an evidence chain; the evidence chain including a snapshot of the original data on which the calculation is based; performing conflict resolution processing based on the initial cost allocation ratio and the real-time budget status of each responsible entity to generate an executable cost allocation ratio verified by the budget; and completing the public cost allocation according to the executable cost allocation ratio and the evidence chain.

[0005] In some optional implementations, the target allocation scenario is determined based on the characteristics of the public expense invoices to be allocated, including: extracting features from the public expense invoices to be allocated to obtain invoice feature information; and identifying the target allocation scenario based on the above invoice feature information through rule matching or classification models. The target allocation scenario includes a fixed-ratio allocation scenario, a motivation-driven allocation scenario, or an experience-based allocation scenario.

[0006] In some optional implementations, based on the aforementioned target cost-sharing scenario and related raw data, the corresponding cost-sharing strategy model is invoked to calculate the initial cost-sharing ratio among the various responsible entities. This includes: obtaining target type input data matching the aforementioned target cost-sharing scenario as the aforementioned related raw data; inputting the aforementioned target type input data into the corresponding cost-sharing strategy model for calculation; wherein, the aforementioned cost-sharing strategy model includes: rule-driven model, usage-driven model, and machine learning model; the corresponding target type input data includes: cost-sharing parameter data determined based on contract terms or preset fixed rules, external system resource consumption data, and historical cost-sharing document data.

[0007] In some optional implementations, the usage-driven model described above is configured to perform the following operations: interface with at least one external resource management system via an application programming interface to obtain detailed resource consumption data associated with each undertaking entity; analyze the tag information contained in the detailed resource consumption data for identifying cost attribution; aggregate the resource consumption based on the tag information, and calculate the sharing ratio among the undertaking entities based on the aggregation results.

[0008] In some optional implementations, based on the aforementioned initial cost-sharing ratio and the real-time budget status of each responsible entity, conflict resolution processing is performed, including: calculating the proposed cost-sharing amount for each responsible entity according to the aforementioned initial cost-sharing ratio and total cost; obtaining the real-time budget balance of each responsible entity; comparing the proposed cost-sharing amount of each responsible entity with the corresponding real-time budget balance to perform budget adequacy verification; and triggering and executing conflict resolution processing when the result of the budget adequacy verification indicates that the proposed cost-sharing amount of at least one responsible entity exceeds the corresponding real-time budget balance.

[0009] In some optional implementations, the above conflict resolution process includes: transferring the excess amount to a pre-defined superior budget pool or public budget pool; and / or adjusting the sharing ratio between at least two responsible entities and recalculating the proposed sharing amount for each responsible entity; and / or generating a budget supplement approval request and suspending the current automatic sharing process.

[0010] In some optional implementations, a chain of evidence is constructed synchronously, including: when calculating the initial cost-sharing ratio, taking a snapshot of the dynamic variable data on which the calculation is based, including at least one of the following: the number of people in the organization at the time of calculation, the effective contract terms version, and the resource consumption logs of the external system; associating and storing the data saved in the snapshot, the identifier of the cost-sharing strategy model called, and the key parameters on which the calculation is based based on the cost-sharing strategy model, to generate structured attribution information; and generating readable natural language description text based on the structured attribution information.

[0011] In some optional implementations, after generating the budget-verified executable cost-sharing ratio, the method further includes: sending a recommended scheme containing the executable cost-sharing ratio and the chain of evidence to the user for review and confirmation; receiving the user's manual adjustment instruction for the recommended scheme and the associated adjustment reasons; and storing the user's final confirmed cost-sharing scheme, the manual adjustment instruction, and the adjustment reasons as feedback data for optimizing the cost-sharing strategy model.

[0012] In some optional implementations, the above method also includes: generating accounting vouchers corresponding to this allocation process based on the above executable cost allocation ratio; associating and integrating the above evidence chain generated during this allocation process, the records generated during budget verification and conflict resolution, and the adjustment records from the user end; and archiving and storing the associated and integrated data with the above accounting vouchers to support audit backtracking based on the above accounting vouchers.

[0013] Secondly, embodiments of the present invention provide a public cost allocation processing system, comprising: a scenario determination module, used to determine a target allocation scenario based on the characteristics of the public cost invoice to be allocated; a ratio calculation and evidence chain construction module, used to call the corresponding allocation strategy model according to the target allocation scenario and related original data, calculate the initial cost allocation ratio among the responsible entities, and simultaneously construct an evidence chain; the evidence chain includes a snapshot corresponding to the original data on which the calculation is based; a budget verification module, used to perform conflict resolution processing based on the initial cost allocation ratio and the real-time budget status of each responsible entity to generate an executable cost allocation ratio that has passed budget verification; and a cost allocation module, used to complete the public cost allocation according to the executable cost allocation ratio and the evidence chain.

[0014] This invention provides a method and system for public expense allocation, aiming to solve the technical problems of low efficiency and large errors in existing allocation processes. The method includes: determining the target allocation scenario based on the characteristics of the public expense invoice to be allocated; calculating the initial expense allocation ratio by calling the corresponding allocation strategy model according to the target allocation scenario and relevant original data, and simultaneously constructing an evidence chain containing snapshots of the original data; performing conflict resolution processing based on the initial ratio and the real-time budget status of each responsible entity to generate an executable expense allocation ratio; and finally completing the allocation based on the executable ratio and the evidence chain. This invention achieves automation, interpretability, and traceability of the allocation process through automatic calculation using a multi-strategy model and solidification of the evidence chain. Combined with real-time budget verification and intelligent conflict resolution, it significantly improves the efficiency and accuracy of allocation processing. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a public cost sharing method provided in an embodiment of the present invention; Figure 2 A flowchart illustrating another method for allocating public costs according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a public cost sharing processing system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] In the cost management practices of shared financial institutions and group enterprises, the allocation of public expenses (such as cloud resource fees, rent, utilities, and joint marketing expenses) has long faced challenges. Current common practices mainly rely on manual processing and fixed-rule systems: finance personnel typically use Excel to manually collect data, match rules, and calculate allocations. This method is inefficient and prone to errors due to formula mistakes or asynchronous data sources, and it fails to retain clear calculation evidence for audit traceability. Some enterprises use the fixed-ratio allocation function in Enterprise Resource Planning (ERP) systems, but its rules are often static and rigid, making it difficult to adapt to dynamic consumption scenarios such as cloud resource usage, and unable to respond to special allocation needs at the document level. Furthermore, many enterprises adopt a "collect first, allocate later" reactive processing model, leading to delayed budget control and an inability to allocate expenses to specific responsible parties when they occur, resulting in budget overruns and unclear accountability. Existing technologies generally suffer from low processing efficiency, poor allocation accuracy, lack of transparency, and a lack of auditable evidence.

[0019] Based on this, the present invention provides a method and system for public cost allocation, so as to realize the automation, interpretability and traceability of the allocation process, solve the technical problems of low allocation efficiency and large error in the prior art, and improve the efficiency and accuracy of allocation.

[0020] To facilitate understanding of this embodiment, a method for sharing public costs disclosed in this invention will first be described in detail. (See [link to relevant documentation]). Figure 1 The diagram illustrates a method for allocating public costs, which can be executed by an electronic device and mainly includes the following steps S102 to S108: Step S102: Determine the target allocation scenario based on the characteristics of the public expense invoices to be allocated.

[0021] Among them, the public expense documents to be allocated can be the original financial vouchers used to trigger the automated allocation process, that is, structured or semi-structured business documents that carry the fact that the expense occurred and implicitly contain clues to the allocation attribution. They usually contain electronic images or structured messages, such as invoices, expense reports, payment applications, etc.

[0022] The aforementioned documents correspond to a set of key fields that can be used to identify the nature of expenses, the responsible party, and the motivation for allocation, i.e., allocation clues. These include, but are not limited to, supplier name, expense type code, amount, remarks text, associated contract number, associated purchase order number, and business department identifier. These features together form the initial semantic basis for determining "who should bear the expense and under what logic."

[0023] In one embodiment, determining the target allocation scenario based on the characteristics of the public expense invoices to be allocated may include: extracting features from the public expense invoices to obtain invoice feature information; and determining the target allocation scenario based on the invoice feature information through rule matching or a classification model. The target allocation scenario may include a fixed-ratio allocation scenario, a motivation-driven allocation scenario, or an experience-based allocation scenario.

[0024] This method corresponds to the deterministic level of the allocation basis: fixed proportion allocation scenarios rely on preset rigid rules (such as contract terms and master data configuration), motivation-driven allocation scenarios rely on real-time measurable business consumption indicators (such as resource usage and service duration), and experience-based allocation scenarios rely on the statistical summarization of historical behavior patterns (such as the usual allocation method of similar expenses).

[0025] As a concrete example, the method of extracting features from public expense invoices to obtain document feature information can include: parsing the document image or text using Optical Character Recognition (OCR) and Natural Language Processing (NLP) technologies, and extracting structured fields using a predefined entity recognition rule base to obtain document feature information, such as: supplier, expense type, amount, remarks, associated contract number, etc. Among these, unstructured text such as project code and department name parsed from the remarks field can serve as key allocation clues in subsequent scenario determination.

[0026] Specifically, the rule matching or classification model can be a deterministic decision tree based on a business rule engine, or a lightweight classification model (such as a random forest) trained by supervised learning. Both take document feature information as input and output the target allocation scenario label and its confidence level. When the document feature clearly points to a certain type of rule (such as the existence of an associated contract number and complete terms), the rule matching path is used first. When the feature is ambiguous or ambiguous (such as the remarks contain multiple items but there is no contract support), the classification model makes a comprehensive judgment.

[0027] Furthermore, methods for determining the target allocation scenario through rule matching or classification model identification may include: matching document feature information with a preset scenario judgment rule set layer by layer, or inputting it into a trained classification model to generate a probability distribution, and selecting the scenario with the highest probability as the target allocation scenario.

[0028] The identification result can directly determine which type of allocation strategy model (rule-driven / usage-driven / machine learning) to call subsequently, and constrain the original data types to be accessed (such as contract term snapshots for fixed-ratio scenarios and external system resource consumption data for motivation-driven scenarios), thereby realizing dynamic binding and closed-loop linkage of scenarios, data, and models.

[0029] Step S104: Based on the target cost-sharing scenario and relevant raw data, call the corresponding cost-sharing strategy model to calculate the initial cost-sharing ratio among the responsible parties and simultaneously construct the evidence chain.

[0030] The aforementioned chain of evidence includes snapshots of the original data on which the apportionment ratio is calculated. This is a verifiable, structured data set that supports ex post attribution and natural language interpretation. It consists of dynamic variable data on which the calculation is based, the identifier of the apportionment strategy model called, and the key parameters on which the model is executed. This chain of evidence provides an immutable anchor point for subsequent budget verification, human intervention, and audit backtracking.

[0031] Each entity responsible for costs can be the smallest accounting unit that assigns cost liability, i.e., an organizational entity or business unit that assumes the obligation to allocate costs and is subject to budget control, such as a cost center, project number, department code, profit center, or business line. Typically, each entity has a unique identifier in the system and is associated with real-time budget balances, organizational structure relationships, and historical allocation records, thereby ensuring that the allocation results are both executable and traceable.

[0032] In one embodiment, the method of calculating the initial cost-sharing ratio among the various entities by calling the corresponding cost-sharing strategy model based on the target cost-sharing scenario and related raw data may include: obtaining target type input data that matches the target cost-sharing scenario as related raw data; and inputting the target type input data into the corresponding cost-sharing strategy model for calculation.

[0033] The aforementioned allocation strategy models include: rule-driven model, usage-driven model, and machine learning model; the corresponding target type input data includes: allocation parameter data determined based on contract terms or preset fixed rules, external system resource consumption data, and historical allocation document data.

[0034] The three types of models described above form a one-to-one mapping relationship with three types of data: rule-driven models rely on static, deterministic parameters (such as contractually agreed proportions); usage-driven models rely on dynamic, observable business driver data (such as CPU hours and number of printed pages); and machine learning models rely on historical allocation behavior sequences (such as the allocation patterns of similar expenses in similar contexts). Together, they cover the deterministic, dynamic, and empirical dimensions of allocation decisions.

[0035] Furthermore, rule-driven models can be determined based on contract terms or preset fixed rules. For example, they can parse explicitly agreed-upon apportionment clauses in the structured text of related contracts (such as "Party A bears 60%, Party B bears 40%)", or look up rigid apportionment factors configured in the master data system, such as department area ratios and account quotas. Apportionment parameters can include: contract ID and clause number, effective date, contracting party identifier, preset ratio value, applicable scope constraints, etc. These parameters together constitute the complete context for rule execution, ensuring that the apportionment logic is reproducible and verifiable.

[0036] Usage-driven models can be determined based on real-time resource consumption data obtained from external systems, such as IT operations and maintenance systems, administrative equipment management systems, or cloud platform billing interfaces, to obtain resource usage details strongly associated with each responsible entity. The core of this model is to transform physical / logical resource consumption into cost attribution weights.

[0037] Machine learning models can analyze and predict allocation based on historical allocation data. For example, using allocation clues from historical documents as feature vectors and actual allocation schemes as labels, collaborative filtering or time-series classification models can be trained to output probabilistic allocation suggestions that conform to business practices. This model generally does not replace rules, but rather provides reasonable data-driven initial values ​​when rules are missing or ambiguous.

[0038] As a concrete example, the usage-driven model described above can be configured to perform the following operations: connect to at least one external resource management system via an application programming interface to obtain detailed resource consumption data associated with each responsible entity; parse the tag information contained in the detailed resource consumption data to identify cost attribution; aggregate resource consumption based on the tag information, and calculate the sharing ratio among the responsible entities based on the aggregation results.

[0039] Specifically, external resource management systems can be cloud service provider billing platforms (such as AWS CostExplorer), IT asset monitoring systems (such as Zabbix), office equipment management systems (such as Print Audit), or vehicle dispatch systems. Detailed resource consumption data can include: timestamps, resource instance IDs, usage values ​​(such as CPU hours, GB storage, page count), and embedded cost attribution tags.

[0040] Furthermore, the detailed data on resource consumption is analyzed to obtain tag information used to identify cost attribution. This tag information can be Project Tag or Cost Center Tag in the cloud platform, or metadata fields such as department code and project number bound to the device system. This tag is the key semantic bridge that maps the original usage to the responsible entity.

[0041] Preferably, the above data parsing methods may include regular expression matching, key-value pair extraction, or lightweight JSON / XMLSchema validation to ensure the robustness of tag recognition.

[0042] Furthermore, the method of aggregating resource consumption based on tag information can include: grouping and summing by tag value to generate a total resource consumption vector for each undertaking entity, and further generating an aggregation result including the consumption proportion of each undertaking entity, standardization coefficient, and abnormal consumption markers; then calculating the sharing ratio among the undertaking entities based on the aggregation result. This ratio is the initial allocation solution corresponding to the consumption-driven model, and its numerical accuracy directly depends on the tag coverage and the granularity of consumption collection.

[0043] As another concrete example, the rule-driven model described above can be configured to perform the following operations: load structured clause fragments of related contracts, extract the subject of the apportionment obligation, the proportion value, and the conditions for effectiveness; simultaneously verify the current validity status of the contract (such as whether it is within the performance period or whether it has been revised); if the verification passes, the proportion agreed in the clause is directly set as the initial apportionment proportion.

[0044] Specifically, this process must ensure that: the contract clause version number is consistent with the effective version registered in the HR / legal system at the time of calculation, and that the scope of application of the clause covers the expense type and occurrence period of this document. This is the premise for the reliable execution of the rule-driven model.

[0045] As another concrete example, the machine learning model described above can be configured to perform the following operations: vectorize the allocation clues of the current document, input them into a trained XGBoost or LightGBM classifier, and output the probability distribution of each responsible entity; normalize and threshold the distribution to generate an initial allocation ratio matrix.

[0046] Specifically, the model input features include: supplier industry classification, expense type coding, amount quantile range, TF-IDF vector of remarks keywords, existence marker of related contracts, and semantic similarity score with historical similar documents. This design enables the model to capture both explicit rules and model implicit business practices, achieving hybrid decision-making.

[0047] In one embodiment, the method of synchronously constructing the chain of evidence may include: when calculating the initial cost-sharing ratio, taking a snapshot of the dynamic variable data on which the calculation is based, the dynamic variable data including at least one of the number of people in the organization at the time of calculation, the effective contract terms version, and the resource consumption log of the external system; associating and storing the snapshot data, the identifier of the cost-sharing strategy model called, and the key parameters on which the calculation is based based on the cost-sharing strategy model, to generate structured attribution information; and generating readable natural language description text based on the structured attribution information.

[0048] Dynamic variable data refers to the basic input data that is timely, variable, and directly affects the accuracy of the results at the time of the amortization calculation; that is, data source instances that may be updated over time or as business status changes. Saving these data as snapshots aims to solidify the true context in which the calculation occurs, preventing the results from becoming unreproducible due to subsequent data changes.

[0049] The identifier of the amortization strategy model being invoked refers to metadata that uniquely identifies the model type, version number, and configuration path used in this instance, namely the Model ID (such as "UsageBased_v2.1_AWS-Tag" or "ContractRule_CNT-202301_Clause3.2"), which can be used to accurately trace the decision engine.

[0050] Key parameters refer to the input variables and their values ​​that play a decisive role in the model execution process, namely the core fields that actually participate in the calculation (such as the contractually agreed proportion of 60%, the CPU hours of a certain project of 12,547 hours, and the historical similar document matching score of 0.92), rather than all the original data, so as to ensure the conciseness and interpretability of the evidence chain.

[0051] Specifically, the data snapshots corresponding to the original data of different allocation strategy models may include: the referenced contract clause ID (i.e., the unique record number in the contract structure database, such as CNT-202301), the snapshot of the usage data source used (i.e., the timestamp section of the original JSON / XML response body obtained when calling the external system API, including cost attribution tag details and usage values), the referenced historical document ID (i.e., the allocation identifier number of the Top-K similar documents used for machine learning inference in the historical database), and the organizational structure snapshot on which this calculation depends (i.e., the snapshot of the number of employees / establishment in the department exported from the "calculation time" in the HR system, such as 53 people in Department A and 47 people in Department B).

[0052] The aforementioned snapshots collectively constitute the objective data set upon which the division is based, and their completeness directly determines whether the chain of evidence meets the audit verifiability requirements. It is important to emphasize that this chain of evidence is not a static log, but rather a directed, structured graph with the allocation event as the root node and each element linked by a unique allocation identifier. It supports both forward tracing (from accounting vouchers, allocation schemes, and the models used to the original data snapshots) and reverse verification (automatically marking historical allocation records that depend on the current number of employees in the HR system as "recalculation required"), thus ensuring that the lifecycle of evidence is manageable, traceable, and interconnected.

[0053] Step S106: Based on the initial cost sharing ratio and the real-time budget status of each responsible entity, perform conflict resolution processing to generate an executable cost sharing ratio that has been verified by the budget.

[0054] The real-time budget status of each responsible entity refers to the dynamic financial data, or budget water level, returned in real time by the budget management system during allocation calculations. This data reflects the current available budget. Typically, it includes key fields such as total budget, used amount, remaining balance, frozen amount, and budget effective period, collectively forming the objective basis for budget adequacy verification. This real-time nature ensures consistency between the verification results and actual business operations, avoiding overspending and missed detections due to T+1 synchronization delays.

[0055] In one embodiment, conflict resolution processing is performed based on the initial cost-sharing ratio and the real-time budget status of each responsible entity, including: calculating the proposed cost-sharing amount for each responsible entity according to the initial cost-sharing ratio and the total cost; obtaining the real-time budget balance of each responsible entity; comparing the proposed cost-sharing amount of each responsible entity with the corresponding real-time budget balance to perform budget adequacy verification; and triggering and performing conflict resolution processing when the result of the budget adequacy verification indicates that the proposed cost-sharing amount of at least one responsible entity exceeds the corresponding real-time budget balance.

[0056] The proposed allocation amount, which is the product of the initial allocation ratio and the total cost, is the direct input for budget verification. The real-time budget balance is the net available amount returned by the budget system at the time of calculation, which is not locked by other processes. The budget adequacy verification achieved by comparing these two is essentially a pre-emptive resource feasibility verification. Its purpose is not to negate the allocation logic itself, but to ensure that the allocation plan has immediate execution capability from a financial perspective. This signifies that this invention upgrades the traditional ex-post allocation and ex-post accountability model to a proactive budget governance paradigm of in-process allocation and in-process risk control.

[0057] In one embodiment, the above conflict resolution process may include any of the following: (1) Transfer the excess amount to the pre-set superior budget pool or public budget pool; This is the Overflow strategy, which is suitable for group-based management scenarios. It allows excess amounts that cannot be covered by a single entity to be automatically transferred to a higher-level responsibility center (such as a department, business unit, or group) or a shared budget pool across departments. This ensures timely accounting and balance of accounts for the responsible entity without modifying the original allocation logic.

[0058] (2) Adjust the apportionment ratio between at least two responsible entities and recalculate the proposed apportionment amount for each responsible entity; This is the rescaling strategy, which is suitable for scenarios with flexible negotiation space (such as joint marketing expenses and shared service fees). The system dynamically reconstructs the allocation ratio matrix based on preset rules (such as weighted by historical contribution or inversely proportional to budget surplus) or user interaction guidance, so as to achieve overall budget balance under the premise that the total amount remains unchanged.

[0059] (3) Generate a budget increase approval request and suspend the current automatic allocation process.

[0060] This is a strong control and blocking strategy. This strategy is applicable to rigid budget control scenarios (such as special funds and special project budgets). When there is no backup budget pool and the proportion cannot be adjusted, the system automatically triggers a structured approval flow (including budget gap analysis, assignment of responsible persons, and time limit warning) and sets the allocation process to "pending approval" status until approval is granted or the process is downgraded due to timeout, ensuring that budget discipline is inviolable.

[0061] It is important to emphasize that the strategies for resolving the above three types of conflicts are not mutually exclusive options, but can form a hierarchical response mechanism: the system prioritizes low-intervention strategies (such as overflow), and only escalates to higher levels (to heavy scaling, and finally to strong control blocking) when the strategy conditions are not met.

[0062] The design logic of this mechanism is that the essence of conflict resolution is not to cover up budget contradictions, but to make financial constraints explicit, structured, and traceable and integrate them into the allocation decision-making closed loop: each overflow transfer records the target budget pool ID and the transfer amount; each proportional rescaling generates a new and old proportional comparison matrix and adjustment basis (such as "because the budget balance of project A is lower than the threshold of 5%, it is rescaled according to the remaining amount"); each strong control blockade solidifies the approval request number and the associated voucher number.

[0063] Furthermore, all data in the above process can be included in the evidence chain archiving scope described in step S108 as records generated during budget verification and conflict resolution, thereby achieving the dual goals of budget control and traceable allocation.

[0064] Step S108: Based on the executable cost sharing ratio and the chain of evidence, complete the sharing of public costs.

[0065] The executable cost-sharing ratio refers to the allocation scheme generated after conflict resolution in step S106, which has passed the budget adequacy verification. Essentially, it is a legally binding allocation instruction with financial enforceability. The chain of evidence refers to the structured attribution set, constructed concurrently in step S104, containing dynamic variable snapshots, model identifiers, and key parameters. Essentially, it is a complete factual basis system supporting the legality of the instruction. Together, these two elements constitute the dual-element foundation for cost-sharing execution; the former defines how much to allocate, and the latter proves the basis for the allocation.

[0066] Completing the allocation of public expenses is not merely the end of index value calculation; rather, it refers to transforming the allocation result into a deterministic financial act with accounting, legal, and auditing validity. This involves generating accounting vouchers that comply with accounting standards and establishing an immutable two-way binding relationship between them and the evidence chain, thus formally elevating the allocation act from an internal system calculation to a formal financial event for the enterprise.

[0067] In another embodiment, after generating the executable cost-sharing ratio that has been budget-verified, the method may further include: sending a recommended scheme containing the executable cost-sharing ratio and the chain of evidence to the user for review and confirmation; receiving a manual adjustment instruction from the user for the recommended scheme and the associated reasons for adjustment; and storing the final cost-sharing scheme, the manual adjustment instruction, and the reasons for adjustment confirmed by the user as feedback data for optimizing the cost-sharing strategy model.

[0068] In this embodiment, the user terminal refers to the terminal interface used by authorized financial auditors, business managers, or budget administrators who participate in the cost-sharing decision-making process; that is, the interactive entry point for human-machine collaborative governance.

[0069] The recommended solution refers to the optimal solution generated by the system based on an algorithm, but its output format must meet the triple requirements of readability, comparability, and operability. Specifically, readability is reflected in the structured allocation matrix (the rows of the matrix represent the responsible parties, and the columns represent the amount / proportion / basis summary); comparability is reflected in the highlighted prompts indicating differences from historical allocation schemes, budget level trends, and similar document allocation patterns; and operability is reflected in the controlled editing capabilities that support fine-tuning of proportions, adding or deleting responsible parties, and mandatory filling in of reasons.

[0070] User-initiated adjustments, which are business-side corrections to the algorithm's output, along with their associated reasons (such as "business consensus," "special policy exemption," or "temporary data source anomaly"), constitute high-quality feedback signals for model optimization. This design allows human experience to be injected into the machine learning loop in a structured manner. The reasons users input for their manual adjustments provide effective training feedback for the allocation strategy model. This feedback data, after being structured, is used to periodically update model parameters, making subsequent allocation recommendations more aligned with actual business decision-making habits.

[0071] In another embodiment, the above method may further include: generating accounting vouchers corresponding to this allocation process based on the executable cost allocation ratio; associating and integrating the evidence chain generated during this allocation process, the records generated during budget verification and conflict resolution, and the adjustment records from the user terminal; and archiving and storing the associated and integrated data with the accounting vouchers to support audit backtracking based on the accounting vouchers.

[0072] In this embodiment, accounting vouchers refer to standardized financial vouchers that meet the requirements of enterprise accounting standards, namely, statutory accounting records with a multi-debit and multi-credit structure, whose core fields (summary, account, amount, auxiliary accounting items) must be strictly consistent with the executable allocation ratio.

[0073] Association integration refers to using the unique identifier of the allocation event (Allocation ID) as the root node to construct a logically consistent traceability graph from heterogeneous data such as voucher data, evidence chain snapshots, budget verification logs, conflict resolution paths, and manual adjustment traces.

[0074] Associated archiving storage refers to writing the graph into a dedicated audit database in an immutable manner (such as hash anchoring or timestamp storage), and embedding the Allocation ID in the "Reference Field" or "Attachment Link" of the ERP voucher. The resulting audit view starts from the accounting voucher, allowing auditors to access a complete picture of the entire allocation process with just one voucher number. This includes original documents, allocation basis, budget verification results, automatic system adjustment records, and manual modification traces and reasons. It truly achieves a closed loop of full-link interpretability from result to motivation, from facts to basis, and from static voucher to dynamic process.

[0075] Based on the same inventive concept, this invention also provides a public cost sharing system in its embodiments, see [link to relevant documentation]. Figure 2 As shown, the system mainly includes the following parts: The scenario determination module 210 is used to determine the target allocation scenario based on the characteristics of the public expense invoice to be allocated; The proportion calculation and evidence chain construction module 220 is used to call the corresponding cost sharing strategy model based on the target cost sharing scenario and relevant original data, calculate the initial cost sharing ratio between each responsible party, and simultaneously construct the evidence chain; the evidence chain includes snapshots of the original data on which the calculation is based. The budget verification module 230 is used to perform conflict resolution processing based on the initial cost allocation ratio and the real-time budget status of each responsible entity, so as to generate an executable cost allocation ratio that has been verified by the budget. The cost allocation module 240 is used to complete the allocation of public costs based on the executable cost allocation ratio and the chain of evidence.

[0076] The public cost sharing processing system provided in this embodiment of the invention can be specific hardware on a device or software or firmware installed on the device. The system provided in this embodiment of the invention has the same implementation principle and technical effects as the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the system embodiments can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.

[0077] As a preferred example, a public cost sharing processing system provided by an embodiment of the present invention may include: The document feature extraction and allocation scenario recognition module is used to receive expense documents and extract document features using optical character recognition (OCR) and natural language processing (NLP) technologies; and automatically recognize allocation scenarios based on document features.

[0078] The sharing ratio calculation and recommendation module is used to retrieve the corresponding sharing strategy model based on the identified sharing scenario, and to calculate the sharing ratio and recommend sharing schemes.

[0079] The budget adequacy verification and conflict resolution module is used to calculate the budget level of each responsible entity based on the recommended allocation scheme. If the budget of a responsible entity is insufficient to pay the allocated amount, conflict resolution is triggered, including automatically finding a backup budget pool, prompting to adjust the ratio, or triggering an over-budget approval process, to ensure the feasibility of the allocation scheme.

[0080] The interpretability evidence chain construction module is used to record a snapshot of the relevant original data on which the calculation is based while calculating the allocation scheme, and to generate the corresponding natural language explanation using a template (i.e., generate readable structured natural language description text).

[0081] Specifically, snapshots of the relevant raw data may include: the ID of the contract clause referenced, snapshots of the usage data source used, the ID of the historical document referenced, budget check results, etc.

[0082] The interaction module is used to send the recommended solution and its corresponding natural language explanation to the client, presenting it to the relevant reviewers for their confirmation or adjustment.

[0083] Preferably, the interaction module can be configured to perform the following steps: The recommended scheme, which includes the executable cost-sharing ratio and the chain of evidence, is sent to the user for review and confirmation; the user's manual adjustment instructions for the recommended scheme and the associated reasons for the adjustment are received; the user's final confirmed cost-sharing scheme, manual adjustment instructions and reasons for adjustment are stored as feedback data for optimizing the cost-sharing strategy model.

[0084] The execution and traceability archiving module is used to generate multiple accounting entries and write them into the ERP system. It also packages and archives the above allocation calculation process, the data snapshots on which the calculation is based, and manual modification records for subsequent audit backtracking and internal settlement.

[0085] Preferably, the execution and traceability archiving module can be configured to perform the following steps: Based on the executable cost allocation ratio, generate the corresponding accounting voucher for this allocation process; link and integrate the evidence chain generated during this allocation process, the records generated during budget verification and conflict resolution, and the adjustment records from the user end; and archive and store the linked and integrated data with the accounting voucher to support audit backtracking based on the accounting voucher.

[0086] The embodiments of the present invention construct an automatic cost allocation and explainable backtracking system through the above-described system, realizing refined cost accounting and improved financial closing efficiency. The evidence chain snapshot technology greatly reduces the risk of compliance audits. The allocation and budget verification are completed at the document entry stage, and overspending behavior is intercepted in real time, realizing budget control in advance and reducing friction between various parts of the enterprise.

[0087] To facilitate understanding, this embodiment of the invention also provides an application example of a public cost sharing processing method. Next, in conjunction with the above-mentioned preferred public cost sharing processing system, a preferred embodiment of the method will be described in detail.

[0088] See Figure 3 The flowchart shown is another method for allocating public costs. This method mainly includes the following steps S301 to S306: Step S301: Determine the target allocation scenario based on the characteristics of the public expense invoices to be allocated; This step is used to achieve Figure 1 The specific execution of step S102 in the method embodiment shown: "Determine the target allocation scenario based on the characteristics of the public expense invoice to be allocated".

[0089] The system receives structured or semi-structured documents to be allocated, including but not limited to electronic invoices, expense reports, and payment applications; it parses the document images or text using optical character recognition (OCR) and natural language processing (NLP) technologies, and extracts document feature information by combining them with a predefined entity recognition rule base.

[0090] The document feature information, namely the "allocation clue" defined in the above embodiments, includes supplier name, expense type code, amount level, remarks text, associated contract number, associated purchase order number, and business department identifier.

[0091] The notes can be unstructured semantic fields such as project code (e.g., “Proj-A, Proj-B”) and department name (e.g., “R&D Department 1, Department 2”) extracted after NLP parsing, which can be used as key allocation clues for subsequent judgment.

[0092] Based on the aforementioned document feature information, the system identifies the target allocation scenario through rule matching or classification models. The target allocation scenario refers to the "fixed ratio allocation scenario, motivation-driven allocation scenario, or experience-based allocation scenario" defined in the above embodiments: when the document features clearly point to a preset rigid rule (such as the existence of an associated contract number and complete terms), it is matched as a fixed ratio allocation scenario; when the features reflect measurable business consumption motivations (such as the remarks containing "cloud server" and no contract support), it is identified as a motivation-driven allocation scenario; when the features are ambiguous or polysemous (such as "department dinner" or "joint marketing expenses"), it is comprehensively judged by the classification model as an experience-based allocation scenario.

[0093] The identification result can directly determine the type of the sharing strategy model to be called subsequently and the range of raw data to be accessed, forming the basis for the dynamic binding of the target sharing scenario with the relevant raw data and sharing strategy model.

[0094] Step S302: Calculate the initial cost sharing ratio by calling the cost sharing strategy model based on the target cost sharing scenario and relevant raw data, and simultaneously construct a preliminary evidence chain. This step is used to achieve Figure 1 The specific execution of step S104 in the method embodiment shown: "Based on the target cost-sharing scenario and related original data, call the corresponding cost-sharing strategy model, calculate the initial cost-sharing ratio among the responsible entities, and simultaneously construct the evidence chain".

[0095] Based on the target allocation scenario identified by S301, the system obtains the target type input data that matches it as the relevant raw data.

[0096] Specifically, for fixed-ratio sharing scenarios, the sharing parameter data determined based on contract terms or preset fixed rules is obtained, namely, "contract ID and clause number, effective date, contracting party identifier, preset ratio value" as defined in the above embodiments.

[0097] For the motivation-driven allocation scenario, the resource consumption data of external systems is obtained, namely, the "resource consumption details data obtained by connecting to at least one external resource management system through the application programming interface" as defined in the above embodiment. Specifically, this includes logs returned by cloud platform billing (such as AWS Cost Explorer), IT operation and maintenance system (such as Zabbix), or office equipment management system (such as PrintAudit) containing timestamps, resource instance IDs, usage values ​​(such as CPU hours, number of prints), and cost attribution tags (ResourceTag).

[0098] For experience-based allocation scenarios, historical allocation document data is obtained, namely the "historical allocation document data" defined in the above embodiments, for machine learning model inference.

[0099] The system inputs the target types of data into the corresponding allocation strategy models for calculation: the rule-driven model parses the structured contract text and extracts the agreed proportions; the usage-driven model parses the ResourceTag and aggregates the usage by tag to generate the allocation proportions; the machine learning model takes the document feature vector as input and outputs probabilistic allocation suggestions.

[0100] During the calculation process, the system simultaneously initiates the evidence chain construction process, which is the pre-operation of "synchronous construction of evidence chain" in the above embodiment. This includes taking snapshots of the dynamic variable data on which the calculation is based (such as the department head snapshot at the "calculation time" in the HR system, the text snapshot of "CNT-202301 Clause 3.2" in the contract database, and the timestamp section of the cloud bill API response body), and generating a draft of structured attribution information, laying the data foundation for the generation of the complete evidence chain of S304.

[0101] Step S303: Based on the initial cost allocation ratio and the real-time budget status of each responsible entity, perform conflict resolution processing to generate an executable cost allocation ratio that has been verified by the budget. This step is used to achieve Figure 1 In the method embodiment shown, step S106 is a complete execution path: "Based on the initial cost sharing ratio and the real-time budget status of each responsible entity, perform conflict resolution processing to generate an executable cost sharing ratio that has been verified by the budget."

[0102] First, the system calculates the proposed allocation amount for each responsible entity based on the initial cost allocation ratio and total cost output by S302; each responsible entity is the "cost center, project number, department code and other smallest accounting unit" defined in the above embodiment, and its unique identity is linked with the budget management system in real time.

[0103] The system obtains the real-time budget balance of each responsible entity at the calculation point through the standard interface of the budget management system. The real-time budget balance is the core field of the "real-time budget status" defined in the above embodiment, which includes the total budget, the amount already used, the remaining balance, the frozen amount, and the budget effective period. The system compares the amount to be allocated with the real-time budget balance and performs a budget adequacy check.

[0104] When the verification result indicates that the proposed allocation amount of at least one undertaking entity exceeds its real-time budget balance, conflict resolution processing is triggered. The conflict resolution processing is defined in the above embodiments as "transferring the excess allocation amount to a preset superior budget pool or public budget pool" (overflow strategy), "adjusting the allocation ratio between at least two undertaking entities and recalculating" (proportion rescaling strategy), or "generating a budget supplement approval request and suspending the current automatic allocation process" (strong control blocking strategy).

[0105] The above three types of strategies constitute a hierarchical response mechanism: low-intervention spillover strategies are prioritized; scaling up to proportional rescaling is only implemented when the spillover conditions are not met; and finally, strong control blocking is implemented under rigid budget constraints.

[0106] All conflict resolution operations generate structured records, including the overflow target budget pool ID and the transfer amount, the new and old ratio matrix of proportional rescaling and the adjustment basis (such as "because the budget balance of project A is lower than the threshold of 5%, it is rescaled according to the remaining amount"), and the approval request number associated with the strong control blocking. The above records will be included in the scope of association and integration of S306 as "records generated during budget verification and conflict resolution in the above embodiments".

[0107] Step S304: Simultaneously construct a complete chain of evidence including structured attribution and natural language description; This step is used to achieve Figure 1 The specific implementation of "synchronously constructing the chain of evidence" in the method embodiment shown is illustrated.

[0108] Throughout the entire process of calculating the initial ratio in S302 and resolving conflicts in S303, the system continuously collects and solidifies all calculation data: it saves snapshots of dynamic variable data, which are the "number of people in the organization at the calculation time, the effective contract terms version, and the resource consumption log of the external system" as defined in the above embodiment.

[0109] Specifically, this includes snapshots of department headcount exported from the HR system (e.g., "Department A: 53 people, Department B: 47 people"), snapshots of the effective text of "CNT-202301 Clause 3.2" in the contract database, and snapshots of the original JSON containing the ResourceTag in the cloud billing API response body.

[0110] The snapshot data, the identifier of the invoked attribution strategy model (such as “ContractRule_CNT-202301_Clause3.2” or “UsageBased_v2.1_AWS-Tag”), and the key parameters on which the model execution is based (such as the contractually agreed proportion of 60% and the CPU hours of a certain project of 12,547 hours) are associated and stored to generate structured attribution information.

[0111] Based on this attribution information, the system uses Natural Language Generation (NLG) technology to transform structured data into readable natural language description text, namely the "readable natural language description text" defined in the above embodiment, for example: "According to Article 3.1 of the framework agreement CNT-202301, and after verifying the department head ratio displayed by the HR system at the time of calculation (53 people in Department A, 47 people in Department B), the recommended allocation ratio is: 40% for Project A and 60% for Project B."

[0112] The evidence chain uses the AllocationID (unique identifier of the allocated event) as the root node to form a directed structured graph that supports forward tracing and reverse verification. This satisfies all the functional requirements of the evidence chain as a verifiable structured data set that supports ex post facto attribution and natural language interpretation in the above embodiments.

[0113] Step S305: Send the recommended solution and evidence chain to the user for review and confirmation, and receive manual adjustment instructions and feedback on reasons; This step is used to achieve Figure 1 The method embodiment shown in the figure is a human-machine collaborative closed loop that "sends a recommended scheme containing an executable cost-sharing ratio and a chain of evidence to the user for review and confirmation; and receives the user's manual adjustment instructions for the recommended scheme and the associated reasons for the adjustment".

[0114] The system pushes the executable cost allocation ratio generated by S303 and verified by budget, and the complete evidence chain generated by S304, to the user terminal in the form of a structured allocation matrix. The user terminal is the "terminal interface used by financial auditors, business managers or budget administrators who are authorized to participate in the allocation decision" as defined in the above embodiment.

[0115] The allocation matrix consists of rows representing the responsible parties and columns representing the amount, proportion, and summary of the basis. It also supports highlighting differences (such as highlighting in red when the deviation from the previous month's allocation proportion exceeds a threshold) to meet the requirements of readability, comparability, and operability.

[0116] Users can accept system recommendations or perform controlled editing operations, including fine-tuning the ratio and adding or deleting responsible entities. If a user makes a significant modification to the system's recommended ratio (such as changing the AI-recommended 50:50 to 70:30), the system will force a pop-up window to ask for the reason for the adjustment. The reason for the adjustment is the "reason for adjustment" defined in the above embodiment, which includes structured options such as "business consensus", "exemption from special policies", and "temporary abnormality of data source".

[0117] The user's final confirmed cost-sharing plan, manual adjustment instructions, and reasons for adjustment can be stored as high-quality feedback data in the correction database. This data is used to periodically optimize the "cost-sharing strategy model" defined in the above embodiments, especially to improve the fitting accuracy of the machine learning model to business practices, forming a continuous evolutionary closed loop of algorithm recommendation, manual intervention, feedback learning, and model iteration.

[0118] Step S306: Generate accounting vouchers based on the executable cost allocation ratio, and link, integrate, and archive the evidence chain, budget verification and conflict resolution records, and manual adjustment records. This step is used to achieve Figure 1 The method embodiment shown in the figure is to "complete the allocation of public expenses according to the executable cost allocation ratio and the evidence chain", and to "integrate the evidence chain generated in this allocation process, the records generated in the budget verification and conflict resolution process, and the adjustment records from the user end; and to archive and store the integrated data with the accounting vouchers".

[0119] Based on the executable cost allocation ratio finally confirmed by the S305 user, the system generates standardized accounting vouchers that comply with enterprise accounting standards. These accounting vouchers are the "standardized financial vouchers that comply with accounting standards" defined in the above embodiments. They have a multi-debit and multi-credit structure, and their summaries, accounts, amounts, auxiliary accounting items and allocation ratios are strictly consistent.

[0120] Meanwhile, the system uses the unique identifier of the allocation event (AllocationID) as the root node to associate and integrate the following four types of heterogeneous data: (1) the complete evidence chain constructed in S304; (2) the budget verification log and conflict resolution path record generated in S303; (3) the user manual adjustment instructions and adjustment reasons collected in S305; and (4) the original document images or structured messages on which this allocation is based.

[0121] The data package that has been integrated as described above is linked with the aforementioned accounting vouchers and archived in a dedicated audit database through hash anchoring and timestamp notation, and the AllocationID is embedded in the "Reference Field" or "Attachment Link" of the ERP voucher.

[0122] The resulting audit view, namely the complete implementation of "supporting audit backtracking based on accounting vouchers" defined in the above embodiments, allows auditors to retrieve the entire chain of information for this allocation with a single click by simply entering a voucher number. This includes the original documents, the target allocation scenario identification process, the snapshot of the allocation strategy model and key parameters, the budget verification results, the system's automatic adjustment records, and the traces and reasons for manual modifications. This truly achieves a closed loop of full-chain interpretability, from result to motivation, from facts to evidence, and from static vouchers to dynamic processes.

[0123] This invention achieves the following substantial technical effects by constructing a full-process cost-sharing technology system covering scene recognition, ratio calculation, budget verification, evidence preservation, human-machine collaboration, and execution archiving: First, it significantly improves the level of precision in cost accounting and its ability to match business drivers.

[0124] The system automatically identifies the allocation scenario based on the allocation clues contained in the document to be allocated, such as supplier information, expense type, remarks text and associated contract number. When it is determined to be a motivation-driven scenario, it connects in real time with the IT operation and maintenance system, cloud service billing platform or equipment management system to obtain resource consumption details data with cost attribution tags, and completes usage aggregation and weight calculation based on the tags, thereby generating an allocation ratio that strictly corresponds to the actual business consumption.

[0125] This mechanism ensures that the allocation of highly variable public costs such as cloud resource fees and shared service center operating costs no longer relies on experience-based estimations or static rules, but directly reflects the actual usage intensity of each responsible entity. This fundamentally solves the cost distortion problem caused by traditional allocation methods, improves the objectivity of departmental cost collection and the fairness of performance evaluation, and forms an effective incentive for resource-saving behavior through precise cost feedback.

[0126] Second, significantly shorten the financial settlement cycle and comprehensively improve automated processing capabilities.

[0127] The system initiates full-link computation at the initial stage of document processing: automatically completes optical character recognition and semantic parsing, dynamically matches applicable allocation strategies, calls the corresponding model to complete ratio calculation, synchronously connects to the budget system for real-time water level verification, triggers a hierarchical conflict resolution mechanism based on the verification results, and finally generates multi-debit and multi-credit accounting vouchers that comply with accounting standards.

[0128] The entire process requires no manual intervention in data collection, formula configuration, or result verification, reducing the manual Excel allocation work that originally took several working days to minutes. This effectively supports enterprises in achieving their goal of rapid closing of accounts and reduces the manpower and time costs of month-end closing.

[0129] Third, ensure the verifiability and audit traceability of the allocation process.

[0130] At the same time as each allocation calculation occurs, the system automatically captures and solidifies the key contextual data on which the calculation depends, including the effective version of the contract terms, the organizational structure snapshot at the calculation time in the human resources system, the original usage log section returned by the external resource management system, and the unique identifier and core input parameters of the allocation model called. After the above data is structured and organized, attribution information is generated and further converted into natural language description text to form a complete chain of evidence. This chain of evidence is established with the final generated accounting voucher through a unique allocation event identifier to establish an immutable two-way association.

[0131] In an audit scenario, a single voucher number can be used to fully reconstruct all elements of the process, from the original documents to allocation decisions, budget verification, and manual adjustments, completely eliminating compliance concerns caused by missing evidence or irreproducible processes.

[0132] Fourth, achieve a fundamental shift in budget control from post-event remediation to in-event interception.

[0133] During the allocation scheme generation stage, the system simultaneously obtains the real-time available quota of each responsible entity in the budget management system and compares the proposed allocation amount with the budget balance for each entity. Once an overspending situation is detected, the system immediately executes a response according to the preset strategy: transferring the excess portion to the superior budget pool or public budget pool, or dynamically reallocating the proportion according to the rules such as the inverse ratio of the surplus while keeping the total amount unchanged, or automatically suspending the process and initiating a structured budget supplement approval under rigid constraints.

[0134] This mechanism embeds budget feasibility verification into the core of the allocation decision-making process, ensuring that financial resources are confirmed before each expense is recorded, thus avoiding problems such as budget overdrafts, unclear responsibilities, and difficulties in tracing when allocating expenses at the end of the month.

[0135] Fifth, it effectively reduces cross-departmental collaboration friction and enhances the transparency and consensus basis for cost allocation decisions.

[0136] The system provides reviewers with a structured apportionment matrix interface, clearly listing the suggested proportions for each responsible entity, a summary of the calculation basis, and a natural language explanation. It also automatically highlights allocation results that deviate significantly from historical averages. Users can make minor adjustments to the proportions or add or delete entities within a controlled range, and all manual interventions must be accompanied by a structured reason for the adjustment. This reason, along with the operation record, is included in the evidence chain archive.

[0137] The resulting cost allocation report combines technical rigor with business readability, enabling business departments to accurately understand the logical starting point and adjustment motivation of cost allocation. This reduces liability disputes caused by information asymmetry and promotes the evolution of the company's internal cost governance system towards standardization, visualization, and negotiation.

[0138] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, specifically, the electronic device includes a processor and a storage device; the storage device stores a computer program, and the computer program, when run by the processor, executes the method described in any of the above embodiments.

[0139] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 400 includes: a processor 410, a memory 420, a communication interface 430, and a bus 440. The memory 420 stores machine-readable instructions that can be executed by the processor 410. When the electronic device is running, the processor 410 communicates with the memory 420 through the bus 440. The processor 410 executes the machine-readable instructions to perform the steps of the method described above.

[0140] Specifically, the memory 420 and processor 410 can be general-purpose memory and processor, without any specific limitations. When the processor 410 runs the computer program stored in the memory 420, it can execute the above method.

[0141] Processor 410 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 410 or by instructions in software form. The processor 410 may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 420, and processor 410 reads the information from memory 420 and, in conjunction with its hardware, completes the steps of the above method.

[0142] Corresponding to the above method, this embodiment of the invention also provides a computer-readable storage medium storing machine-executable instructions. When the computer-executable instructions are called and run by a processor, the computer-executable instructions cause the processor to perform the steps of the above method.

[0143] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0144] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0145] Furthermore, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0146] It should be noted that if the functionality is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0147] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0148] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for sharing public costs, characterized in that, include: Based on the characteristics of the public expense invoices to be allocated, the target allocation scenario is determined; Based on the target cost-sharing scenario and relevant raw data, the corresponding cost-sharing strategy model is invoked to calculate the initial cost-sharing ratio among the responsible parties, and a chain of evidence is constructed simultaneously; the chain of evidence includes snapshots of the raw data on which the calculation is based. Based on the initial cost-sharing ratio and the real-time budget status of each responsible entity, conflict resolution is performed to generate an executable cost-sharing ratio that has been verified by the budget. The public cost allocation is completed based on the executable cost sharing ratio and the chain of evidence.

2. The method according to claim 1, characterized in that, Based on the characteristics of the public expense invoices to be allocated, the target allocation scenarios are determined, including: Feature extraction is performed on the public expense invoices to be allocated to obtain invoice feature information; Based on the document feature information, the target allocation scenario is determined by rule matching or classification model identification; wherein, the target allocation scenario includes a fixed ratio allocation scenario, a motivation-driven allocation scenario, or an experience-based allocation scenario.

3. The method according to claim 1, characterized in that, Based on the target cost-sharing scenario and relevant raw data, the corresponding cost-sharing strategy model is invoked to calculate the initial cost-sharing ratio among the responsible entities, including: Obtain target type input data that matches the target allocation scenario as the relevant raw data; The target type input data is input into the corresponding amortization strategy model for calculation; The allocation strategy model includes: rule-driven model, usage-driven model and machine learning model; the corresponding target type input data includes: allocation parameter data determined based on contract terms or preset fixed rules, external system resource consumption data, and historical allocation document data.

4. The method according to claim 3, characterized in that, The usage-driven model is configured to perform the following operations: By connecting to at least one external resource management system through an application programming interface (API), detailed data on resource consumption associated with each responsible entity can be obtained. Parse the tag information contained in the resource consumption details data, which is used to identify cost attribution; Based on the tag information, the resource consumption is aggregated, and the sharing ratio among the responsible entities is calculated according to the aggregation results.

5. The method according to claim 1, characterized in that, Based on the initial cost-sharing ratio and the real-time budget status of each responsible party, conflict resolution processing is performed, including: Based on the initial cost-sharing ratio and the total cost, calculate the proposed cost-sharing amount for each responsible party; Obtain the real-time budget balance of each responsible entity; The proposed allocation amount for each responsible entity is compared with the corresponding real-time budget balance to verify the budget adequacy. When the result of the budget adequacy verification indicates that the proposed amount to be shared by at least one undertaking entity exceeds the corresponding real-time budget balance, conflict resolution processing is triggered and executed.

6. The method according to claim 5, characterized in that, The conflict resolution process includes: The excess amount will be transferred to a pre-designated superior budget pool or public budget pool. And / or, adjust the sharing ratio between at least two responsible parties and recalculate the proposed share amount for each responsible party; And / or, generate a budget supplement approval request and pause the current automatic allocation process.

7. The method according to claim 1, characterized in that, Simultaneously construct the chain of evidence, including: When calculating the initial cost sharing ratio, a snapshot of the dynamic variable data on which the calculation is based is saved. The dynamic variable data includes at least one of the following: the number of people in the organization at the time of calculation, the effective version of the contract terms, and the resource consumption log of the external system. The data saved in the snapshot, the identifier of the amortization strategy model called, and the key parameters on which the calculation is based based on the amortization strategy model are associated and stored to generate structured attribution information; Based on the structured attribution information, readable natural language descriptive text is generated.

8. The method according to claim 1, characterized in that, After the step of generating a budget-validated executable cost allocation ratio, the method further includes: The recommended scheme, which includes the executable cost sharing ratio and the chain of evidence, will be sent to the user for review and confirmation. Receive manual adjustment instructions from the user terminal for the recommended scheme and the associated reasons for the adjustment; The user-confirmed cost-sharing scheme, the manual adjustment instruction, and the adjustment reason are stored as feedback data to optimize the cost-sharing strategy model.

9. The method according to claim 1, characterized in that, The method further includes: Based on the executable cost allocation ratio, generate the corresponding accounting voucher for this allocation process; The evidence chain generated during this cost allocation process, the records generated during budget verification and conflict resolution, and the adjustment records from the user end will be linked and integrated. The integrated data is linked to and archived with the accounting vouchers to support audit backtracking based on the accounting vouchers.

10. A public cost sharing system, characterized in that, include: The scenario determination module is used to determine the target allocation scenario based on the characteristics of the public expense invoices to be allocated; The proportion calculation and evidence chain construction module is used to call the corresponding cost sharing strategy model based on the target cost sharing scenario and related original data, calculate the initial cost sharing ratio among the responsible parties, and simultaneously construct the evidence chain; the evidence chain includes a snapshot of the original data on which the calculation is based; The budget verification module is used to perform conflict resolution processing based on the initial cost allocation ratio and the real-time budget status of each responsible entity, so as to generate an executable cost allocation ratio that has been verified by the budget. The cost-sharing module is used to complete the sharing of public costs based on the executable cost-sharing ratio and the evidence chain.

Images