Method and system for identifying abnormal metering material accounting based on knowledge graph and rule center

Abstract

The application relates to the computer technical field and provides a metered material accounting exception identification method and system based on a knowledge graph and a rule center, which comprises the following steps: determining state change information based on entity state transition tracks of an accounting process of metered materials in a knowledge graph, determining a target accounting node based on state change information of an accounting stage; performing rule conflict analysis based on node state information of the target accounting node and associated business rules of the target accounting node in a rule topology network, obtaining an abnormal accounting node; obtaining a candidate accounting node based on a state transition path before the abnormal accounting node according to the entity state transition tracks, performing accounting exception analysis based on node states between the abnormal accounting node and the candidate accounting node and between the candidate accounting nodes, and determining a potential starting node; and determining an accounting exception reason of the metered materials based on an abnormal operation mode of the potential starting node. The application improves the accuracy and reliability of accounting exception identification of the metered materials.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method and system for identifying anomalies in the accounting of measured materials based on knowledge graphs and rule centers. Background Technology

[0002] In the operation of enterprises or institutions, the reimbursement of measured materials is an important and complex task. With the expansion of business scale and the continuous increase in the types and quantities of measured materials, traditional manual review methods for reimbursement are inefficient, error-prone, and unable to meet actual needs. Therefore, methods for identifying anomalies in measured material reimbursement have emerged. Most existing methods for identifying anomalies in measured material reimbursement are based on historical data and simple rules. For example, they identify anomalies by setting fixed numerical thresholds to determine if the reimbursement amount exceeds the threshold, or by using a simple correspondence between material category and reimbursement frequency for preliminary screening. However, existing methods lack in-depth analysis and understanding of the complex relationships between various stages of the measured material reimbursement process. Relying solely on a single dimension or simple rules fails to accurately identify hidden and interconnected anomalies, leading to numerous false alarms or omissions, thus reducing the accuracy and reliability of reimbursement anomaly identification. Summary of the Invention

[0003] This invention provides a method and system for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers, aiming to improve the accuracy and reliability of identifying anomalies in the reimbursement of measured materials.

[0004] In a first aspect, the present invention provides a method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers, including:

[0005] Based on the entity state transition trajectory of the reimbursement process of measured materials in a pre-constructed knowledge graph, the state change information of the measured materials at each reimbursement stage is determined, and the target reimbursement node in the reimbursement process is determined based on the state change information at each reimbursement stage.

[0006] Based on the node status information of each target billing node and its associated business rules in the preset rule topology network, rule conflict analysis is performed to identify abnormal billing nodes.

[0007] Based on the entity state transition trajectory, the state transition path before the abnormal expense reporting node is traced back to obtain candidate expense reporting nodes in the tracing process. Based on the node states between the abnormal expense reporting node and each candidate expense reporting node, as well as between each candidate expense reporting node, expense reporting anomaly analysis is performed to determine the potential starting node that causes the rule conflict of the abnormal expense reporting node.

[0008] The cause of the reimbursement anomaly for the measured materials is determined based on the abnormal operation mode of the potential starting node.

[0009] Secondly, the present invention also provides a measurement material reimbursement anomaly identification system based on knowledge graphs and rule centers, applied to the measurement material reimbursement anomaly identification method based on knowledge graphs and rule centers as described in the first aspect; the measurement material reimbursement anomaly identification system based on knowledge graphs and rule centers includes:

[0010] The reimbursement node analysis module is used to determine the status change information of the measured materials at each reimbursement stage based on the entity state change trajectory of the reimbursement process in a pre-built knowledge graph, and to determine the target reimbursement node in the reimbursement process based on the status change information at each reimbursement stage.

[0011] The rule conflict analysis module is used to perform rule conflict analysis based on the node status information of each target billing node and its associated business rules in the preset rule topology network, and to identify abnormal billing nodes.

[0012] The expense reporting anomaly analysis module is used to backtrack the state transition path before the abnormal expense reporting node based on the entity state transition trajectory, obtain candidate expense reporting nodes in the backtracking process, and perform expense reporting anomaly analysis based on the node states between the abnormal expense reporting node and each candidate expense reporting node, as well as between each candidate expense reporting node, to determine the potential starting node that causes the rule conflict of the abnormal expense reporting node.

[0013] The reimbursement anomaly analysis module is used to determine the reimbursement anomaly causes for measured materials based on the abnormal operation mode of the potential starting node.

[0014] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing computer software programs; and a processor for reading and executing the computer software programs, thereby realizing the above-described method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers.

[0015] Fourthly, the present invention also provides a non-transitory computer-readable storage medium storing a computer software program, which, when executed by a processor, implements the above-described method for identifying anomalies in the accounting of measured materials based on knowledge graphs and rule centers.

[0016] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for identifying anomalies in the accounting of measured materials based on knowledge graphs and rule centers.

[0017] The method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers provided in this invention performs rule conflict analysis between the node state information of the target reimbursement node and the associated business rules in the rule topology network. It marks the abnormal reimbursement nodes with rule conflicts, initially locating the possible location of the anomaly. By tracing back the state transition paths of the abnormal reimbursement nodes with rule conflicts, it identifies the potential starting nodes that caused the rule conflicts, narrowing down the scope of the anomalies. Finally, it determines the cause of the reimbursement anomaly based on the abnormal operation mode of the potential starting nodes. Therefore, it can deeply analyze the complex relationships between each stage of the entire reimbursement process for measured materials. Based on the entity state transition trajectory of the knowledge graph, the rule topology network, and node relationships, it comprehensively and accurately identifies abnormal situations in the reimbursement process, from judging abnormal reimbursement nodes and tracing back anomalies to finally clarifying the cause of the anomaly, thus improving the accuracy and reliability of identifying anomalies in the reimbursement of measured materials. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for identifying anomalies in the reimbursement of measurement materials based on knowledge graphs and rule centers, as provided in an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of the structure of the measurement material reimbursement anomaly identification system based on knowledge graph and rule center provided in an embodiment of the present invention;

[0020] Figure 3 An embodiment diagram of the electronic device provided in this invention;

[0021] Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0023] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0024] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0025] Optional, see below Figure 1 , Figure 1 This is a flowchart illustrating the method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers provided by this invention. In this embodiment, the executing entity of the method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers is the reimbursement management system. Therefore, the method for identifying anomalies in the reimbursement of measured materials based on knowledge graphs and rule centers includes:

[0026] Step 10: Based on the entity state change trajectory of the reimbursement process of the measured materials in the pre-constructed knowledge graph, determine the state change information of the measured materials at each reimbursement stage, and determine the target reimbursement node in the reimbursement process based on the state change information at each reimbursement stage.

[0027] Optionally, the reimbursement management system pre-constructs a knowledge graph. This knowledge graph contains all entities (such as measuring equipment, invoices, approvers, etc.) in the reimbursement process for measured materials, their relationships, and the state transition trajectories of these entities throughout the process. Therefore, in response to the reimbursement process for measured materials, the management system traverses the entity state transition trajectories within the knowledge graph to obtain state change information for the measured materials at each reimbursement stage. For each reimbursement stage, it identifies the state of the entities related to the measured material reimbursement at the beginning of that stage (i.e., the initial reimbursement state) and the state of the entities related to the measured material reimbursement at the end of that stage (i.e., the final reimbursement state). State change information includes key details such as whether the measuring equipment has completed verification, whether the invoice has been submitted, and whether the approval has been granted.

[0028] Furthermore, the expense reimbursement management system determines the target expense reimbursement node from all expense reimbursement stages based on preset screening conditions and status change information of each expense reimbursement stage, as in steps 101 to 104. The target expense reimbursement node may be an expense reimbursement stage involving large-amount payments or requiring approval from multiple departments.

[0029] Step 20: Based on the node status information of each target billing node and its associated business rules in the preset rule topology network, perform rule conflict analysis to obtain abnormal billing nodes.

[0030] Furthermore, the expense reimbursement management system pre-stores a rule topology network. This network contains various business rules related to the expense reimbursement process, defining the conditions and operational procedures that should be met at different node states. Therefore, for each target expense reimbursement node, the system obtains its node status information. This information includes all key status data of the target node during the expense reimbursement process. The system then compares the node status information of each target node with the associated business rules in the rule topology network. If the actual status information of the target expense reimbursement node does not match the status conditions specified in the associated business rules (i.e., a rule conflict occurs), the target expense reimbursement node is identified as an abnormal node.

[0031] Continuing with the electricity company's meter procurement and reimbursement process, let's assume the target reimbursement nodes are "Contract Signing," "Delivery and Acceptance," and "Financial Review." For the "Contract Signing" node, its status information is "Contract signed successfully, terms meet requirements," which aligns with the relevant business rule in the rule topology network: "Contracts must be legally reviewed and terms must be without loopholes," indicating no rule conflict. The status information for the "Delivery and Acceptance" node is "Acceptance passed, report submitted," which also satisfies the rule "Acceptance must be performed according to standard procedures and results must be genuine," again showing no conflict.

[0032] The "Financial Audit" target reimbursement node's status information is "Invoice amount does not match contract amount, and no explanation for the discrepancy is provided." In the rule topology network, the business rules associated with the "Financial Audit" node clearly stipulate that "the invoice amount must match the contract amount; if there is a discrepancy, a reasonable explanation must be provided." After comparison, the reimbursement management system found that this node's status did not meet the business rule requirements, indicating a rule conflict. Therefore, the "Financial Audit" node was determined to be an abnormal reimbursement node.

[0033] Step 30: Based on the entity state transition trajectory, backtrack the state transition path before the abnormal reimbursement node to obtain the candidate reimbursement nodes in the backtracking process. Then, based on the node states between the abnormal reimbursement node and each candidate reimbursement node, as well as between each candidate reimbursement node, perform reimbursement anomaly analysis to determine the potential starting node that causes the rule conflict of the abnormal reimbursement node.

[0034] Furthermore, the expense reimbursement management system, based on the entity state transition trajectories in the knowledge graph, traces back the state transition paths before the abnormal expense reimbursement nodes. During the tracing process, all nodes traversed along the path are recorded; these nodes are the candidate expense reimbursement nodes. Continuing with the above embodiment, in the electricity company's meter procurement expense reimbursement process, "financial review" has been identified as an abnormal expense reimbursement node. Tracing back its previous state transition paths yields candidate expense reimbursement nodes such as "delivery acceptance," "contract signing," "supplier selection," and "purchase application."

[0035] Furthermore, the expense reimbursement management system performs expense reimbursement anomaly analysis on the node status between the abnormal expense reimbursement node and each candidate expense reimbursement node, as well as the node status among each candidate expense reimbursement node. By comparing the logical relationships and data changes between different node states, it determines whether there are unreasonable state transitions or data anomalies. If a state change of a candidate expense reimbursement node causes rule conflicts in subsequent nodes, then that candidate expense reimbursement node is identified as a potential starting node causing rule conflicts with the abnormal expense reimbursement node, as detailed in steps 301 to 304.

[0036] Step 40 determines the cause of the reimbursement anomaly for the measured materials based on the abnormal operation mode of the potential starting node.

[0037] Furthermore, the reimbursement management system conducts a detailed analysis of the operation records and related data of potential starting nodes to identify abnormal operation patterns. These abnormal operation patterns include non-standard operation procedures, data entry errors, and improper use of permissions. Further, the reimbursement management system analyzes these abnormal operation patterns based on the business logic of the reimbursement process for measured materials to determine the reasons for the reimbursement anomalies, as detailed in steps 401 to 404.

[0038] This invention analyzes the node status information of the target reimbursement node against the associated business rules in the rule topology network to identify abnormal reimbursement nodes with rule conflicts. It initially locates the possible location of the anomaly. By tracing back the state transition paths of the abnormal reimbursement nodes with rule conflicts, it identifies the potential starting node that caused the rule conflict, narrowing down the scope of the anomaly. Finally, it determines the cause of the reimbursement anomaly for the measured materials based on the abnormal operation mode of the potential starting node. Therefore, it can deeply analyze the complex relationships between each stage of the reimbursement process for measured materials. Based on the entity state transition trajectory of the knowledge graph, the rule topology network, and node relationships, it comprehensively and accurately identifies abnormal situations in the reimbursement process, from judging abnormal reimbursement nodes and tracing back anomalies to finally clarifying the cause of the anomaly. This improves the accuracy and reliability of identifying reimbursement anomalies for measured materials.

[0039] In one embodiment, steps 101 to 104 include:

[0040] Step 101: Perform business logic analysis based on the start and end states of each expense reporting stage to obtain the business logic relationships between each expense reporting stage, and construct a business logic relationship diagram for each expense reporting stage based on the business logic relationships between each expense reporting stage.

[0041] Optionally, the expense reimbursement management system performs business logic analysis on the start and end states of each expense reimbursement stage. By comparing the start and end states of different stages, it identifies business logic relationships such as dependencies (e.g., the end state of the previous stage is the start state of the next stage) and causal relationships (e.g., an operation in the previous stage directly leads to a change in the state of the next stage). Furthermore, the expense reimbursement management system presents these business logic relationships graphically, constructing a business logic relationship diagram for each expense reimbursement stage, where nodes represent expense reimbursement stages and edges represent the business logic relationships between stages.

[0042] Continuing with the above analysis of the meter procurement and reimbursement process, we find that the "procurement application" stage ends with "procurement application approved," while the "supplier selection" stage begins, indicating a direct dependency between the two. Furthermore, the "contract signing" stage ends with "contract effective," which is one of the starting conditions for the "delivery and acceptance" stage, demonstrating a causal relationship. Therefore, the business logic diagram for the reimbursement stages can be constructed as follows: edges connect nodes such as "procurement application" → "supplier selection," "contract signing" → "delivery and acceptance," forming a complete business logic network.

[0043] Step 102: Based on the business logic relationship diagram of the reimbursement stage, determine the target stage where each reimbursement stage is related and the degree of close business logic relationship is greater than a preset threshold, and determine the core reimbursement stage based on the number of target stages for each reimbursement stage.

[0044] Furthermore, based on the business logic relationship diagram of the expense reimbursement stages, the expense reimbursement management system calculates the degree of business logic correlation between each expense reimbursement stage and other stages. This correlation is determined by a combination of the number of direct connections between expense reimbursement stages and the strength of these connections (e.g., strong connections that are mandatory, weak connections that are optional). When the correlation of a certain expense reimbursement stage exceeds a preset threshold, the associated stage is identified as the target stage for that expense reimbursement stage.

[0045] Furthermore, the expense reimbursement management system counts the target number of each expense reimbursement stage and identifies the top 30% of stages by target number as core expense reimbursement stages.

[0046] Continuing with the above embodiments, the preset threshold is "the number of target stages ≥ 2 and the proportion of strong correlation ≥ 50%", in the metering procurement and reimbursement process:

[0047] The "financial review" stage is strongly correlated with the entire reimbursement process. Assuming the entire reimbursement process comprises 8 stages, the target number of stages is 7, meeting the threshold. The "goods receipt and acceptance" stage is strongly correlated with the "invoice submission" and "financial review" stages, so the target number of stages is 2, meeting the threshold. The "contract signing" stage is strongly correlated with the "advance payment application," "purchase application," "financial review," "invoice submission," and "goods receipt and acceptance" stages, so the target number of stages is 5, meeting the threshold. Therefore, the reimbursement management system identifies "financial review," "goods receipt and acceptance," and "contract signing" as the core reimbursement stages.

[0048] Step 103: For each core expense reimbursement stage, determine its impact on the entire expense reimbursement process based on the direct triggering or restricting effect of its status change information on the status changes of subsequent expense reimbursement stages, and sort each core expense reimbursement stage according to its impact range to obtain an impact ranking list.

[0049] Furthermore, for each core expense reporting stage, the expense reporting management system analyzes the direct triggering or restricting effect of its status change information on subsequent stages. Direct triggering effect means that the end status of the core stage directly starts the subsequent stage (e.g., "goods received and accepted" directly triggers "invoice submission"); restricting effect means that if the status of the core stage does not meet the conditions, the subsequent stage cannot proceed (e.g., "contract signing not completed" restricts the start of "goods received and accepted").

[0050] Furthermore, the expense management system determines the scope of impact for each core stage based on the number of subsequent stages affected and the degree of criticality of the impact (e.g., the impact weight is higher when it involves funds), and sorts the core stages from largest to smallest in terms of impact scope, generating an impact ranking list.

[0051] Continuing with the above examples, an analysis of the core reimbursement stages "financial review", "goods receipt and acceptance", and "contract signing" reveals that changes in the status of "financial review" directly affect and limit the final completion of the entire process, with the impact covering the entire reimbursement stage. Assuming the entire reimbursement stage includes 8 stages;

[0052] The change in the "goods arrival and acceptance" status directly triggers "invoice submission" and restricts the initiation of "financial audit", affecting two stages (including one core stage).

[0053] The change in the "contract signing" status directly triggers "advance payment application", "material procurement / service order placement" and "financial review", and restricts the prerequisites for starting subsequent stages such as "delivery acceptance" and "invoice submission" (subsequent procurement acceptance and financial related processes cannot be carried out without a signed and effective contract). The impact covers 5 stages (including core stages such as project execution, material procurement, and financial review).

[0054] Therefore, the impact ranking list is as follows: 1. Financial audit; 2. Contract signing; 3. Goods arrival and acceptance.

[0055] Step 104: Based on the business logic relationship diagram and the impact ranking list of the expense reimbursement stage, perform potential bottleneck analysis to obtain the target expense reimbursement node.

[0056] Furthermore, the expense reimbursement management system performs a bottleneck potential analysis based on the business logic relationship diagram and the impact ranking list of the expense reimbursement stage to obtain the target expense reimbursement node, as detailed in steps 1041 to 1044.

[0057] This invention accurately identifies key reimbursement nodes that have a critical impact on the overall process from a complex reimbursement process, providing precise monitoring targets for subsequent rule conflict analysis and anomaly tracing. This enables precise management of the reimbursement process for measured materials, thereby improving the accuracy and reliability of identifying reimbursement anomalies for measured materials.

[0058] In one embodiment, steps 1041 to 1044 include:

[0059] Step 1041: Based on the position and connection information of the target reimbursement stage (first preset position in the impact sorting list) in the business logic relationship graph, determine whether each target reimbursement stage is at a bottleneck position in the business logic. Whether it is at a bottleneck position in the business logic indicates whether the number of subsequent reimbursement stages that cannot proceed normally after a problem occurs in the target reimbursement stage is greater than or equal to a preset threshold.

[0060] Optionally, the expense reimbursement management system obtains the target expense reimbursement stage that ranks first among the top preset positions in the impact ranking list; that is, it selects the expense reimbursement stage that ranks first among the top preset positions in the impact ranking list. For each target expense reimbursement stage, the management system extracts its location information and connection information from the expense reimbursement stage's business logic relationship graph. The location information refers to the stage's node position in the relationship graph, and the connection information is the number and type of edges connecting it to other expense reimbursement stages. Further, the management system calculates the number of subsequent expense reimbursement stages that would be unable to proceed normally if a problem occurs in the target expense reimbursement stage, and compares this number with a preset threshold to determine if the target expense reimbursement stage is at a bottleneck position in the business logic. If the number of affected subsequent expense reimbursement stages is greater than or equal to the preset threshold, the target expense reimbursement stage is considered to be at a bottleneck position; otherwise, it is not at a bottleneck position.

[0061] Continuing with the above embodiment, the first preset number of digits is 3, the top 3 target reimbursement stages in the impact sorting list are "financial review", "contract signing" and "goods arrival and acceptance", and the preset quantity threshold is 2.

[0062] Financial review stage: In the business logic relationship diagram, "financial review" is directly connected to the entire expense reimbursement stage. Assuming that the entire expense reimbursement stage includes 8 stages, the number of subsequent expense reimbursement stages affected is 7, which is greater than the preset threshold. Therefore, "financial review" is at the bottleneck position of the business logic.

[0063] Goods arrival and acceptance stage: In the business logic diagram, "Goods arrival and acceptance" is directly connected to "Invoice submission" and "Financial review". If a problem occurs in the "Goods arrival and acceptance" stage (such as acceptance failure), it will directly cause "Invoice submission" and "Financial review" to be unable to proceed normally. The number of subsequent reimbursement stages affected is 2, which is equal to the preset quantity threshold. Therefore, "Goods arrival and acceptance" is at the bottleneck position of the business logic.

[0064] Contract signing stage: "Contract signing" is directly linked to "advance payment application", "material procurement / service order placement", "financial review", "invoice submission" and "delivery acceptance". If there is a problem with "contract signing" (such as the contract not taking effect), "advance payment application", "material procurement / service order placement", "financial review", "delivery acceptance" and "invoice submission" will not be able to proceed normally. The number of subsequent reimbursement stages affected is 5, which is greater than the preset threshold. Therefore, "contract signing" is at the bottleneck position of the business logic.

[0065] Step 1042: Identify the target billing stage that is at the bottleneck position of the business logic as the potential bottleneck billing stage.

[0066] Furthermore, the expense reimbursement management system directly identifies the target expense reimbursement stage, which is at a bottleneck in the business logic, as a potential bottleneck expense reimbursement stage. If a problem occurs at a potential bottleneck expense reimbursement stage, it will significantly impact multiple subsequent stages. Continuing with the above example, "financial review," "contract signing," and "goods receipt and acceptance" are identified as potential bottleneck expense reimbursement stages because these three stages are at bottlenecks in the business logic; problems at these stages will prevent at least two subsequent expense reimbursement stages from proceeding normally.

[0067] Step 1043: Based on the number and complexity of the association paths that are logically related to other reimbursement stages in the reimbursement stage business logic relationship diagram for each potential bottleneck reimbursement stage, determine the stage importance of each potential bottleneck reimbursement stage.

[0068] Furthermore, for each potential bottleneck reimbursement stage, the reimbursement management system counts the number of logically related paths between that stage and other reimbursement stages in the business logic relationship diagram. A higher number of related paths indicates more frequent interactions and greater importance for that stage within the process. The system further assesses the complexity of these relationships, determined by the connection type of the related paths (e.g., unidirectional dependency, bidirectional dependency, multiple-choice dependency, etc.), with more complex connection types like bidirectional or multiple-choice dependencies receiving higher weights. Finally, by combining the number of related paths and the complexity of the relationships, the system determines the stage importance of each potential bottleneck reimbursement stage; a higher number of related paths and greater complexity indicate higher stage importance.

[0069] Continuing with the aforementioned bottlenecks in the reimbursement process, namely "financial review," "contract signing," and "goods receipt and inspection":

[0070] Financial review stage: There are 7 related paths, which connect the entire expense reporting stage, and all of them are bidirectional dependencies.

[0071] Contract signing stage: There are 5 associated paths, connecting "Prepayment Application", "Material Procurement / Service Order Placement", "Financial Review", "Goods Acceptance", and "Invoice Submission". Among them, "Prepayment Application", "Material Procurement / Service Order Placement", and "Goods Acceptance" have a one-way dependency relationship, "Financial Review" has a two-way dependency relationship, and "Invoice Submission" has an indirect relationship (contract signing affects goods acceptance, and goods acceptance affects invoice submission).

[0072] Goods arrival and acceptance stage: There are 2 associated paths, connecting "Invoice Submission" and "Financial Review", and both are unidirectional dependencies.

[0073] Based on comprehensive assessment, the "financial review" stage is of higher importance than the "contract signing" stage, and the "contract signing" stage is of higher importance than the "delivery and acceptance" stage.

[0074] Step 1044: Sort the stages based on their importance for each potential bottleneck reimbursement stage to obtain an importance ranking list, and determine the reimbursement stage that ranks second to last in the importance ranking list as the target reimbursement node.

[0075] Furthermore, the expense reimbursement management system sorts the expense reimbursement stages based on their importance, resulting in an importance ranking list. In this embodiment of the invention, the second preset number of digits is 3. Therefore, the expense reimbursement management system selects the top 3 expense reimbursement stages from the importance ranking list and determines them as target expense reimbursement nodes. These target expense reimbursement nodes occupy a critical position in the entire expense reimbursement process.

[0076] Continuing with the above embodiments, in addition to "financial review", "contract signing" and "goods receipt and acceptance", there are also "stage A" and "stage B" in the importance ranking list. The top 3 reimbursement stages are "financial review", "contract signing" and "goods receipt and acceptance". Therefore, the reimbursement management system determines "financial review", "contract signing" and "goods receipt and acceptance" as the target reimbursement nodes.

[0077] This invention, through its embodiments, gradually filters out target reimbursement nodes that are bottlenecks in the business logic and of high importance from the target reimbursement stages with significant impact. Therefore, it can accurately locate the core and critical links in the reimbursement process, thereby enabling key monitoring and anomaly analysis, improving the stability and risk control capabilities of the reimbursement process, and avoiding reimbursement delays or errors caused by anomalies in critical links.

[0078] In one embodiment, steps 301 to 304 include:

[0079] Step 301: Based on the first node state of the abnormal reimbursement node and the second node state of each candidate reimbursement node, determine the difference in the initial node state between each candidate reimbursement node and the abnormal reimbursement node.

[0080] Optionally, the expense reimbursement management system obtains the first node status of the abnormal expense reimbursement node and the second node status of each candidate expense reimbursement node. The first node status and the second node status both contain information such as the node's key business data and process status in the expense reimbursement process.

[0081] Furthermore, the accounting management system compares the second node status of each candidate accounting node with the first node status of the abnormal accounting node field by field to find inconsistencies in data or status between the two, thereby determining the differences in the initial node status between each candidate accounting node and the abnormal accounting node.

[0082] Continuing in the meter purchase reimbursement process, it is known that "financial review" is an abnormal reimbursement node, with its first node status being "the invoice amount is inconsistent with the contract amount, and no explanation of the difference is provided"; "contract signing" and "goods receipt and acceptance" are candidate reimbursement nodes. The second node status of "contract signing" is "the payment amount agreed in the contract is inconsistent with the actual purchase order amount, and no correction has been made", and the second node status of "goods receipt and acceptance" is "goods have been accepted and the acceptance report has been submitted".

[0083] A comparison of the expense reimbursement management system revealed the following:

[0084] Compared to the "financial audit" node, the "contract signing" node has a different initial node status in terms of consistency of amount and handling of discrepancies.

[0085] The "Goods Arrival and Acceptance" node and the "Financial Audit" node have different initial node states in terms of amount-related status, and the "Goods Arrival and Acceptance" node does not involve any prior status information regarding amount anomalies.

[0086] Step 302: Based on the rule topology network, filter the state differences of each initial node to obtain the state differences of the target node that are related to the conflict of business rules in the rule topology network.

[0087] Furthermore, the expense reimbursement management system matches each initial node state difference with the business rules in the rule topology network one by one. The rule topology network stores all the business rules in the expense reimbursement process, including constraints on node states. It determines whether each initial node state difference violates the business rules in the rule topology network. If it does, it is filtered out to obtain the target node state differences that conflict with the business rules in the rule topology network.

[0088] Continuing with the above embodiments, in the rule topology network, the business rules related to meter purchase reimbursement include "the contract amount must be consistent with the purchase order amount" and "the invoice amount must be consistent with the contract amount; if there is a discrepancy, a reasonable explanation must be provided." Regarding the initial node state difference between the "Contract Signing" node and the "Financial Review" node, the discrepancy between the payment amount stipulated in the contract and the actual purchase order amount violates the business rule "the contract amount must be consistent with the purchase order amount." Regarding the initial node state difference between the "Goods Arrival and Acceptance" node and the "Financial Review" node, while the "Goods Arrival and Acceptance" node itself does not involve a direct rule violation, logically, its failure to intervene in contract amount anomalies indirectly conflicts with the rule "the invoice amount must be consistent with the contract amount." The reimbursement management system filters out discrepancies in the "Contract Signing" node amount and potential discrepancies related to the amount rules in the "Goods Arrival and Acceptance" node as target node state differences.

[0089] Step 303: Based on the second node state of every two candidate reimbursement nodes, perform node state correlation analysis to obtain the similarities and differences in node states. Based on the similarities and differences in node states between every two candidate reimbursement nodes, obtain the state relationship matrix between the candidate reimbursement nodes. The matrix elements of the state relationship matrix represent the node state relationship between every two candidate reimbursement nodes.

[0090] Furthermore, the expense reimbursement management system performs node status correlation analysis on the second node status of every two candidate expense reimbursement nodes. By comparing them, it identifies the similarities and differences in node status. The similarities in node status indicate that the two nodes are consistent in certain business data or process statuses, while the differences in node status indicate that the two nodes differ in certain business data or process statuses.

[0091] Furthermore, the expense reimbursement management system constructs a state relationship matrix based on the similarities and differences in the node states between any two candidate expense reimbursement nodes. The state relationship matrix is ​​a two-dimensional matrix, where rows and columns correspond to different candidate expense reimbursement nodes. Matrix elements represent the node state relationship between any two candidate expense reimbursement nodes; for example, "1" indicates a direct association and similar states, "-1" indicates conflicting or logically contradictory state differences, and "0" indicates no obvious association.

[0092] Continuing with the two candidate reimbursement points: "Contract signing" and "Delivery and acceptance":

[0093] Similarities in node status: Neither node effectively corrected or explained the abnormal contract amount situation.

[0094] The differences in node status are as follows: The "Contract Signing" node directly causes the inconsistency in amount, while the "Delivery and Acceptance" node is mainly responsible for the acceptance of goods and does not involve the setting of contract amount.

[0095] The state relationship matrix constructed by the expense reimbursement management system is as follows:

[0096]

[0097] The state relationship matrix shows that the "contract signing" and "delivery and acceptance" nodes are related to the abnormal contract amount issue and have similar states.

[0098] Step 304: Perform an accounting anomaly analysis based on the target node state differences and state relationship matrix to obtain potential starting nodes.

[0099] Furthermore, the expense reimbursement management system performs expense reimbursement anomaly analysis based on the target node status differences and status relationship matrix to obtain potential starting nodes, as detailed in steps 3041 to 3044.

[0100] This invention can start from the status information of abnormal reimbursement nodes and related candidate reimbursement nodes, gradually analyze the differences in node status, filter differences related to business rule conflicts, construct a state relationship matrix between nodes, and finally comprehensively determine the potential starting node that causes the rule conflict of the abnormal reimbursement node. This achieves accurate tracing of reimbursement anomalies, finds the source of problems in the reimbursement process, and improves the accuracy and reliability of identifying reimbursement anomalies of measured materials.

[0101] In one embodiment, steps 3041 to 3044 include:

[0102] Step 3041: Based on the state relationship matrix, determine the target node state relationship where the pattern similarity of the state difference pattern with the target node state difference is greater than the preset similarity threshold, and determine the two candidate accounting nodes corresponding to the target node state relationship as the pattern association candidate node pair.

[0103] Optionally, the expense reimbursement management system defines the state difference pattern of the target node, i.e., the type of data anomaly or state conflict presented by the state difference of the target node. Further, the system traverses each node state relationship in the state relationship matrix and calculates the similarity between each node state relationship and the target node state difference pattern. The similarity calculation is based on dimensions such as the data characteristics of the node state differences and the degree of business logic correlation. When the pattern similarity between a node state relationship and the target node state difference pattern is greater than a preset similarity threshold, that node state relationship is determined as the target node state relationship, and its two corresponding candidate expense reimbursement nodes form a pattern association candidate node pair.

[0104] Continuing with the above embodiment, the known target node state differences are "contract amount inconsistency at the contract signing node" and "contract amount anomaly due to lack of intervention at the delivery and acceptance node." The state difference pattern manifests as anomalies related to contract amount and subsequent nodes failing to address these anomalies. A preset similarity threshold is set to 0.7. The state relationship matrix shows that the state relationship between the "contract signing" and "delivery and acceptance" nodes is 1, indicating a correlation and similarity between the two. Analysis reveals that the "contract signing" node directly generates the inconsistency issue, while the "delivery and acceptance" node fails to correct or explain this problem. The state relationship between these two nodes highly matches the target node state difference pattern in terms of contract amount anomalies and unaddressed issues. The calculated pattern similarity is 0.8, greater than the preset threshold. Therefore, "contract signing" and "delivery and acceptance" form a pattern-related candidate node pair.

[0105] Step 3042: For each first candidate node pair in the pattern association candidate node pair, based on the causal relationship that the state change of the first candidate node in the first candidate node pair causes the subsequent candidate node to have a state change related to the rule conflict, determine the causal rule association candidate node pair in the first candidate node pair.

[0106] Furthermore, for each pattern-associated candidate node pair (i.e., the first candidate node pair), the expense management system analyzes the state changes of the first candidate node in the first candidate node pair to determine whether it will cause subsequent candidate nodes to have state changes related to rule conflicts. Here, the first candidate node is the preceding node in the first candidate node pair, that is, the node that ranks first.

[0107] If such a causal relationship exists, that is, the state change of the first candidate node is the direct or indirect cause of the subsequent node's state violating the business rules, then the first candidate node pair is determined as a causal rule-related candidate node pair, as in steps 30421 to 30424.

[0108] Step 3043: For each second candidate node pair in the causal rule-related candidate node pair, based on the number of subsequent candidate nodes affected by the state transmission caused by the state change of the second candidate node in each second candidate node pair, determine the state transmission influence of each second candidate node, and determine the second candidate node whose state transmission influence is greater than the preset influence threshold as a potential candidate node.

[0109] Furthermore, for each causal rule-related candidate node pair (i.e., the second candidate node pair), the expense management system analyzes the state changes of each second candidate node in the second candidate node pair and counts the number of subsequent candidate nodes affected by the state transmission caused by the state change. The second candidate node is the preceding node in the second candidate node pair, that is, the node ranked first. The influence of state transmission is positively correlated with the number of subsequent candidate nodes affected. The more nodes affected, the greater the influence of state transmission.

[0110] Furthermore, the expense management system compares the status transmission influence of each second candidate node with a preset influence threshold. If the status transmission influence of a second candidate node is greater than the preset influence threshold, then the second candidate node is identified as a potential candidate node.

[0111] In one embodiment, assuming the causal rule-associated candidate node pair is a "contract signing - goods receipt and acceptance" pattern, the change in the "contract signing" node's state affects two subsequent candidate nodes: "goods receipt and acceptance" and "financial audit." The "goods receipt and acceptance" node's state change without intervention only affects one subsequent candidate node: "financial audit." The preset influence threshold is set to affect one subsequent node. The influence propagated by the "contract signing" node's state is 2, which is greater than the preset influence threshold, while the influence propagated by the "goods receipt and acceptance" node's state is 1, which is equal to the preset influence threshold. Therefore, the "contract signing" node is identified as a potential candidate node.

[0112] Step 3044: Based on the matching of the state changes of each potential candidate node in the backtracking path with the business rules in the rule topology network, determine the potential candidate node with the earliest state change that does not conform to the business rules, and determine the earliest inconsistent potential candidate node as the potential starting node.

[0113] Furthermore, the expense reimbursement management system analyzes the matching of the state changes of each potential candidate node with the business rules in the rule topology network in the backtracking path. According to the time sequence of the expense reimbursement process, it checks whether the state changes of potential candidate nodes conform to the business rules in turn, finds the potential candidate node with the earliest state change that does not conform to the business rules, and determines it as the potential starting node. This node is the source node that causes subsequent expense reimbursement anomalies and rule conflicts.

[0114] Continuing with the identified potential candidate nodes for "Contract Signing" and other potential candidate nodes, the "Contract Signing" node exhibited a status change during the contract signing process where "the payment amount stipulated in the contract is inconsistent with the actual purchase order amount, and no correction has been made," violating the business rule that "the contract amount must be consistent with the purchase order amount." In contrast, the status changes of other potential candidate nodes all occurred after the "Contract Signing" node entered an abnormal state, thus the "Contract Signing" node was the earliest potential candidate node to exhibit a status change inconsistent with the business rules and was identified as the potential starting node.

[0115] This invention starts from the differences in the state of the target node and the state relationship matrix. By filtering the node relationships with similar patterns, identifying causal nodes, and assessing the influence of state transmission, it can accurately locate the earliest potential starting node that violates business rules. Therefore, it can systematically analyze the complex relationships and state changes between nodes in the reimbursement process, and improve the accuracy and reliability of identifying reimbursement anomalies of measured materials.

[0116] In one embodiment, steps 30421 to 30424 include:

[0117] Step 30421: Based on the state difference information related to the conflict between the first candidate node and the rule in the first candidate node pair, determine whether the first candidate node meets the conditions for triggering the state change related to the conflict between the subsequent candidate nodes and the rule.

[0118] Optionally, for the first candidate node among the first candidate nodes, the expense management system extracts its status difference information related to rule conflicts. The status difference information includes specific data that violates business rules in the node status, abnormal process status, and other content.

[0119] Furthermore, based on the business logic of the expense reimbursement process and the relevant rules in the rule topology, the expense reimbursement management system determines whether the state difference information of the first candidate node has the potential to cause subsequent candidate nodes to undergo state changes related to rule conflicts. If the state difference information of the first candidate node can logically lead to subsequent nodes violating business rules, then it is determined that it meets the state change condition; otherwise, it does not.

[0120] Continuing with the first candidate node pair in the "Contract Signing - Delivery and Acceptance" process, the status difference information related to rule conflict for the first candidate node "Contract Signing" is "The payment amount stipulated in the contract is inconsistent with the actual purchase order amount, and no correction has been made." In the meter purchase and reimbursement process, business rules require that "the contract amount must be consistent with the purchase order amount," and the "Delivery and Acceptance" stage must be based on contract information for acceptance, while the "Financial Audit" stage must verify the invoice and contract amounts. From a business logic perspective, the inconsistent amount status at the "Contract Signing" node may lead to the subsequent "Delivery and Acceptance" not being able to be accepted according to the correct amount, and may trigger rule conflicts during "Financial Audit" due to the discrepancy between the invoice and contract amounts. Therefore, it is determined that the "Contract Signing" node meets the conditions for triggering the status change related to rule conflicts in subsequent candidate nodes.

[0121] Step 30422: If the state change condition is met, then based on the relationship between the state change of the first candidate node in the first candidate node pair and the business rules, determine the logical mode of rule conflict propagation.

[0122] Furthermore, once the first candidate node meets the state change conditions, the expense reimbursement management system analyzes the relationship between the first candidate node's state change and the business rules, clarifying how the first candidate node's state change propagates the rule conflict to subsequent candidate nodes through each stage of the expense reimbursement process. Based on this propagation process, the logical pattern of rule conflict propagation is determined, including direct causal relationships, indirect causal relationships, and multi-node related causal relationships, clarifying the specific propagation path and method of the conflict from the first candidate node to subsequent candidate nodes.

[0123] Continuing with the first candidate node pair of "Contract Signing - Goods Delivery and Acceptance," the "Contract Signing" node exhibits a status change where "the payment amount stipulated in the contract is inconsistent with the actual purchase order amount, and no correction has been made." Its relationship to the business rules is a direct violation of the rule that "the contract amount must be consistent with the purchase order amount." From the perspective of the reimbursement process, this status change propagates in an indirect causal logic: the inconsistency in the "Contract Signing" node prevents the "Goods Delivery and Acceptance" stage from accepting goods based on the correct contract amount. Although the "Goods Delivery and Acceptance" stage itself does not directly violate the rules, the unprocessed contract amount discrepancy inevitably leads to a situation where "the invoice amount is inconsistent with the contract amount, and no explanation for the discrepancy is provided" when the subsequent "Financial Audit" stage verifies the invoice and contract amounts, triggering a rule conflict that "the invoice amount must be consistent with the contract amount, and a reasonable explanation must be provided if there is a discrepancy." Therefore, the rule conflict propagation logic for this first candidate node pair is determined to be an indirect causal relationship, with the propagation path being "Contract Signing → Goods Delivery and Acceptance → Financial Audit."

[0124] Step 30423: Based on the logical pattern of the first candidate node, determine whether there is a reasonable logical path from the state change of the first candidate node to the state change of subsequent candidate nodes that conflict with the rules.

[0125] Furthermore, based on the logical pattern and combined with the actual business operations and rule requirements of the expense reimbursement process, the expense reimbursement management system determines whether there is a reasonable logical path from the state change of the first candidate node to the state change of subsequent candidate nodes that conflict with the rules. The reasonable logical path requires that in the business process, the state change of the first candidate node can naturally trigger the state change of subsequent candidate nodes that are related to the rule conflict in accordance with the normal operation sequence and rule constraints, without logical contradictions or unreasonable jumps.

[0126] Continuing with the first candidate node pair, "Contract Signing - Goods Acceptance," the rule conflict propagation logic follows an indirect causal relationship, with the propagation path being "Contract Signing → Goods Acceptance → Financial Audit." In the actual business process of meter procurement and reimbursement, after contract signing, the goods acceptance stage begins, based on contract information. After acceptance, the financial audit stage begins, requiring verification of the invoice and contract amounts. The discrepancy in amounts at the "Contract Signing" node perfectly aligns with the normal business operation sequence, naturally transmitting the issue through the goods acceptance stage to the financial audit stage, leading to a rule conflict in the financial audit. There is no logical contradiction. Therefore, it is determined that a reasonable logical path exists from the change in status at the "Contract Signing" node to the change in status of rule conflict at the "Financial Audit" node.

[0127] Step 30424: If a reasonable logical path exists, the first candidate node pair with a reasonable logical path is determined as a candidate node pair for causal rule association.

[0128] Furthermore, when a reasonable logical path is determined to exist from the state change of the first candidate node to the state change of subsequent candidate nodes that conflict with the rules, the expense reimbursement management system identifies the first candidate node pair with this reasonable logical path as a causal rule-related candidate node pair. Continuing with the first candidate node pair "Contract Signing - Goods Acceptance," the expense reimbursement management system determines that a reasonable logical path exists from the state change of the "Contract Signing" node to the state change of the "Financial Audit" node that conflicts with the rules. Therefore, the "Contract Signing - Goods Acceptance" first candidate node pair is identified as a causal rule-related candidate node pair.

[0129] This invention systematically and comprehensively analyzes the causal relationships between nodes by judging state change conditions, determining the logical pattern of rule conflict propagation, and verifying logical paths. This avoids misjudging accidental or logically unreasonable node pairs as key node pairs related to reimbursement anomalies, accurately determines candidate node pairs associated with causal rules, and improves the accuracy and reliability of identifying reimbursement anomalies of measured materials.

[0130] In one embodiment, steps 401 to 404 include:

[0131] Step 401: The abnormal operation mode is broken down into multiple independent expense reimbursement process steps. Based on the time sequence of each expense reimbursement process step, the time sequence relationship between each expense reimbursement process step is obtained.

[0132] Optionally, after the expense reimbursement management system obtains the abnormal operation mode of the potential starting node, it breaks down the abnormal operation mode into multiple independent expense reimbursement process steps according to the actual steps of the expense reimbursement process. Each step corresponds to a specific operation or task in the expense reimbursement process, such as data entry, review, and submission.

[0133] Furthermore, the expense reimbursement management system identifies the chronological order of each step based on their actual occurrence time, thus obtaining the time sequence relationship between each step of the expense reimbursement process.

[0134] Continuing with the meter purchasing and reimbursement process, "contract signing" has been identified as a potential starting point, with its abnormal operation pattern being "the payment amount stipulated in the contract is inconsistent with the actual purchase order amount, and no correction has been made." The reimbursement management system breaks down this abnormal operation pattern into the following reimbursement process steps: ① Contract drafting and amount entry; ② Internal contract review; ③ Contract finalization and signing. Based on the operation times recorded by the system, the time sequence is determined as follows: first, contract drafting and amount entry are performed; then, the internal contract review stage begins; finally, the contract finalization and signing are completed.

[0135] Step 402: Based on the time sequence relationship, determine the associated reimbursement actions between adjacent reimbursement process steps, and perform action standardization analysis on each associated reimbursement action based on the preset reimbursement standard actions to obtain the initial abnormal reimbursement actions that do not conform to the standard.

[0136] Furthermore, the expense reimbursement management system analyzes the associated expense reimbursement actions occurring between adjacent stages of the expense reimbursement process based on time sequence relationships. Associated expense reimbursement actions refer to the operations performed when transitioning from one stage to the next, such as submitting for review or confirming approval. The system further compares each associated expense reimbursement action with preset expense reimbursement standard actions, which outline the standard procedures and requirements that each operation in the expense reimbursement process should follow. If an associated expense reimbursement action does not conform to the standards of the preset expense reimbursement standard actions, it is identified as an initial abnormal expense reimbursement action that does not meet the standards.

[0137] Continuing with the abnormal operation pattern at the potential starting point of "contract signing," we analyzed the related reimbursement actions between adjacent stages: From "contract drafting and amount entry" to "internal contract review," the related reimbursement action is "submitting contract for review." The pre-defined reimbursement procedure requires verifying the consistency between the contract amount and the purchase order amount before submission for review. However, this verification was not performed in practice, making the "submitting contract for review" action non-compliant and thus identified as the initial abnormal reimbursement action. From "internal contract review" to "contract finalization and signing," the related reimbursement action is "approval and confirmation of final draft." Since no discrepancy was found during the review stage, this action is also non-compliant. However, compared to the previous action, "submitting contract for review," the abnormality occurred earlier and is the source of subsequent problems.

[0138] Step 403: Based on the mutual influence relationship between the abnormal reimbursement process links corresponding to the initial abnormal reimbursement action, construct the abnormal link influence relationship network, and analyze the reimbursement problem based on the abnormal link influence relationship network and the business logic of the reimbursement process of the measured materials, and determine the target abnormal reimbursement action that will cause subsequent reimbursement problems.

[0139] Furthermore, the expense reimbursement management system analyzes the interrelationships between the abnormal expense reimbursement process steps corresponding to the initial abnormal expense reimbursement actions, constructing an abnormal step influence relationship network. In this network, nodes represent abnormal expense reimbursement process steps, and edges represent the influence relationships between steps, such as causal relationships and dependency relationships. Further, the system analyzes this network in conjunction with the business logic of the material reimbursement process to determine which initial abnormal expense reimbursement actions will directly or indirectly trigger subsequent expense reimbursement problems. Those initial abnormal expense reimbursement actions that have a critical impact on subsequent expense reimbursement problems and lead to the escalation of abnormalities are identified as target abnormal expense reimbursement actions that trigger these problems.

[0140] Continuing with the abnormal operation mode of "Contract Signing," the network of impact relationships of the abnormal links constructed shows that: An error in the amount entered during the "Contract Drafting and Amount Entry" stage affects the "Internal Contract Review" stage (preventing it from detecting the problem), which in turn affects the "Contract Finalization and Signing" stage (the erroneous contract is signed), ultimately leading to a rule conflict in the "Financial Review" stage where the invoice and contract amounts do not match. Combining this with the business logic of the reimbursement process, the initial abnormal reimbursement action of not verifying the amount when submitting the "Contract Drafting and Amount Entry" stage for review is the key factor leading to a series of subsequent problems; therefore, it is identified as the target abnormal reimbursement action.

[0141] Step 404: Determine the cause of the abnormal expense report based on the target abnormal expense report action.

[0142] Furthermore, the expense reimbursement management system takes the target abnormal expense reimbursement action as the core clue, analyzes its causes and consequences, and, by combining information such as the business background of the expense reimbursement process, operating procedures, and operation records of relevant personnel, clarifies how the target abnormal expense reimbursement action violated regulations and caused subsequent expense reimbursement abnormalities, thus determining the cause of the expense reimbursement abnormality for the measured materials.

[0143] Continuing with the meter purchasing and reimbursement process, the target abnormal reimbursement action was the failure to verify the amount during the "contract drafting and amount entry" stage. The reimbursement management system, through reviewing operation records, discovered that the contract drafter, due to negligence, failed to verify the contract amount against the purchase order amount according to the prescribed procedure, and the reviewer also failed to fulfill their strict review responsibilities. Therefore, the cause of this reimbursement abnormality was determined to be "the contract drafter's failure to verify the contract amount as required, and the ineffective oversight in the review stage, leading to the signing of an incorrect contract and subsequent reimbursement abnormality due to discrepancies between the invoice and contract amount during financial review."

[0144] This invention can start from the abnormal operation mode of potential starting nodes, gradually break down the operation process, analyze the standardization of actions, and construct an influence relationship network to accurately locate the target abnormal reimbursement action that causes the reimbursement abnormality and determine the root cause of the reimbursement abnormality. Therefore, it systematically sorts out the entire process of abnormal operation from generation to problem, providing an accurate basis for subsequent targeted improvement measures, optimization of reimbursement process, and strengthening of personnel management, thereby improving risk prevention and control capabilities.

[0145] Furthermore, the anomaly identification system for reimbursement of measurement materials based on knowledge graph and rule center provided by the present invention will be described below. The anomaly identification system for reimbursement of measurement materials based on knowledge graph and rule center described below can be referred to in correspondence with the anomaly identification method for reimbursement of measurement materials based on knowledge graph and rule center described above.

[0146] Optional, refer to Figure 2 , Figure 2This is a schematic diagram of the structure of the measurement material reimbursement anomaly identification system based on knowledge graph and rule center provided by the present invention. The measurement material reimbursement anomaly identification system based on knowledge graph and rule center includes:

[0147] The reimbursement node analysis module 210 is used to determine the status change information of the measured materials in each reimbursement stage based on the entity status change trajectory of the reimbursement process in the pre-built knowledge graph, and to determine the target reimbursement node in the reimbursement process based on the status change information of each reimbursement stage.

[0148] The rule conflict analysis module 220 is used to perform rule conflict analysis based on the node status information of each target billing node and its associated business rules in the preset rule topology network to obtain abnormal billing nodes.

[0149] The reimbursement anomaly analysis module 230 is used to backtrack the state transition path before the abnormal reimbursement node based on the entity state transition trajectory, obtain the candidate reimbursement nodes in the backtracking process, and perform reimbursement anomaly analysis based on the node state between the abnormal reimbursement node and each candidate reimbursement node, as well as between each candidate reimbursement node, to determine the potential starting node that causes the rule conflict of the abnormal reimbursement node.

[0150] The reimbursement anomaly analysis module 240 is used to determine the reimbursement anomaly causes for measured materials based on the abnormal operation modes of potential starting nodes.

[0151] This invention analyzes the node status information of the target reimbursement node against the associated business rules in the rule topology network to identify abnormal reimbursement nodes with rule conflicts. It initially locates the possible location of the anomaly. By tracing back the state transition paths of the abnormal reimbursement nodes with rule conflicts, it identifies the potential starting node that caused the rule conflict, narrowing down the scope of the anomaly. Finally, it determines the cause of the reimbursement anomaly for the measured materials based on the abnormal operation mode of the potential starting node. Therefore, it can deeply analyze the complex relationships between each stage of the reimbursement process for measured materials. Based on the entity state transition trajectory of the knowledge graph, the rule topology network, and node relationships, it comprehensively and accurately identifies abnormal situations in the reimbursement process, from judging abnormal reimbursement nodes and tracing back anomalies to finally clarifying the cause of the anomaly. This improves the accuracy and reliability of identifying reimbursement anomalies for measured materials.

[0152] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3As shown, this embodiment of the invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it performs the following steps:

[0153] Based on the entity state transition trajectory of the reimbursement process of measured materials in a pre-constructed knowledge graph, the state change information of measured materials at each reimbursement stage is determined, and the target reimbursement node in the reimbursement process is determined based on the state change information at each reimbursement stage.

[0154] Based on the node status information of each target billing node and its associated business rules in the preset rule topology network, rule conflict analysis is performed to identify abnormal billing nodes.

[0155] Based on the entity state transition trajectory, the state transition path before the abnormal reimbursement node is traced back to obtain the candidate reimbursement nodes in the backtracking process. Based on the node state between the abnormal reimbursement node and each candidate reimbursement node, as well as between each candidate reimbursement node, the reimbursement anomaly analysis is performed to determine the potential starting node that causes the rule conflict of the abnormal reimbursement node.

[0156] The cause of accounting anomalies for measured materials is determined based on the abnormal operation patterns of potential starting nodes.

[0157] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it performs the following steps:

[0158] Based on the entity state transition trajectory of the reimbursement process of measured materials in a pre-constructed knowledge graph, the state change information of measured materials at each reimbursement stage is determined, and the target reimbursement node in the reimbursement process is determined based on the state change information at each reimbursement stage.

[0159] Based on the node status information of each target billing node and its associated business rules in the preset rule topology network, rule conflict analysis is performed to identify abnormal billing nodes.

[0160] Based on the entity state transition trajectory, the state transition path before the abnormal reimbursement node is traced back to obtain the candidate reimbursement nodes in the backtracking process. Based on the node state between the abnormal reimbursement node and each candidate reimbursement node, as well as between each candidate reimbursement node, the reimbursement anomaly analysis is performed to determine the potential starting node that causes the rule conflict of the abnormal reimbursement node.

[0161] The cause of accounting anomalies for measured materials is determined based on the abnormal operation patterns of potential starting nodes.

[0162] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the above-described method for identifying anomalies in measurement material reimbursement based on knowledge graphs and rule centers. This method includes:

[0163] Based on the entity state transition trajectory of the reimbursement process of measured materials in a pre-constructed knowledge graph, the state change information of measured materials at each reimbursement stage is determined, and the target reimbursement node in the reimbursement process is determined based on the state change information at each reimbursement stage.

[0164] Based on the node status information of each target billing node and its associated business rules in the preset rule topology network, rule conflict analysis is performed to identify abnormal billing nodes.

[0165] Based on the entity state transition trajectory, the state transition path before the abnormal reimbursement node is traced back to obtain the candidate reimbursement nodes in the backtracking process. Based on the node state between the abnormal reimbursement node and each candidate reimbursement node, as well as between each candidate reimbursement node, the reimbursement anomaly analysis is performed to determine the potential starting node that causes the rule conflict of the abnormal reimbursement node.

[0166] The cause of accounting anomalies for measured materials is determined based on the abnormal operation patterns of potential starting nodes.

[0167] The system embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0168] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A knowledge graph and rule center-based measurement material accounting exception identification method, characterized in that, The method comprises the following steps: Based on the entity state transition track in the pre-constructed knowledge graph, the state change information of the metered materials in each accounting stage is determined, and the target accounting node in the accounting process is determined based on the state change information of each accounting stage; Based on the node state information of each target accounting node and its associated business rules in the preset rule topology network, rule conflict analysis is performed to obtain an abnormal accounting node; Based on the entity state transition track, the state transition path before the abnormal accounting node is traced back to obtain a candidate accounting node in the tracing process, and accounting abnormality analysis is performed based on the node state between the abnormal accounting node and each candidate accounting node and between each candidate accounting node, to determine a potential starting node that causes the rule conflict of the abnormal accounting node; Based on the abnormal operation mode of the potential starting node, the accounting abnormality reason of the metered materials is determined.

2. The knowledge graph and rule center-based metrology material accounting exception identification method of claim 1, wherein, The state change information represents the accounting start state at the beginning of each accounting stage and the accounting end state at the end; the target accounting node in the accounting process is determined based on the state change information of each accounting stage, which comprises: Based on the accounting start state and the accounting end state of each accounting stage, business logic analysis is performed to obtain the business logic relationship between each accounting stage, and an accounting stage business logic relationship graph is constructed based on the business logic relationship between each accounting stage; Based on the accounting stage business logic relationship graph, a target stage with associated and closely related business logic is determined for each accounting stage, and a core accounting stage is determined based on the number of target stages of each accounting stage; For each core accounting stage, the influence range of the entire accounting process is determined according to the direct triggering or limiting effect of its state change information on the state change of the subsequent accounting stage, and each core accounting stage is sorted according to the influence range to obtain an influence sorting list; Based on the accounting stage business logic relationship graph and the influence sorting list, bottleneck potential analysis is performed to obtain the target accounting node.

3. The knowledge graph and rule center-based metrology material accounting exception identification method of claim 2, wherein, The bottleneck potential analysis based on the accounting stage business logic relationship graph and the influence sorting list to obtain the target accounting node comprises: Based on the position information and connection information of the target accounting stage in the first preset number of positions in the influence sorting list in the business logic relationship graph, it is determined whether each target accounting stage is in a bottleneck position of business logic; whether in a bottleneck position of business logic represents whether the number of subsequent accounting stages that cannot normally proceed after the target accounting stage has a problem is greater than or equal to a preset number threshold; The target accounting stage in the bottleneck position of business logic is determined as a bottleneck potential accounting stage; Based on the number of associated paths and the complexity of the associated paths of each bottleneck potential accounting stage in the accounting stage business logic relationship graph with other accounting stages, the stage importance of each bottleneck potential accounting stage is determined; Sort based on the importance of each bottleneck potential billing stage, get the importance ranking list, and determine the billing stage in the importance ranking list as the target billing node.

4. The knowledge graph and rule center-based metrology material accounting exception identification method of claim 1, wherein, The abnormal billing analysis based on the node state between the abnormal billing node and each candidate billing node and between each candidate billing node determines the potential starting node that causes the rule conflict of the abnormal billing node, including: Based on the first node state of the abnormal billing node and the second node state of each candidate billing node, determine the initial node state difference between each candidate billing node and the abnormal billing node; Based on the rule topology network, filter each initial node state difference to obtain a target node state difference related to the business rule conflict in the rule topology network; Based on the second node state of each two candidate billing nodes, perform node state association analysis to obtain node state same points and node state different points, and based on the node state same points and node state different points between each two candidate billing nodes, obtain a state relationship matrix between candidate billing nodes; The matrix elements of the state relationship matrix represent the node state relationship between each two candidate billing nodes; Based on the target node state difference and the state relationship matrix, perform billing exception analysis to obtain the potential starting node.

5. The knowledge graph and rule center-based metrology material accounting exception identification method of claim 4, wherein, The abnormal billing analysis based on the target node state difference and the state relationship matrix to obtain the potential starting node includes: Based on the state relationship matrix, determine a target node state relationship with a state difference mode similar to the target node state difference, and the mode similarity is greater than a preset similarity threshold; and determine two candidate billing nodes corresponding to the target node state relationship as a mode-associated candidate node pair; For each first candidate node pair in the mode-associated candidate node pair, based on the causal relationship that the state change of the first candidate node in the first candidate node pair triggers the state change related to the rule conflict in the subsequent candidate node, determine a causal rule-associated candidate node pair in the first candidate node pair. For each second candidate node pair in the causal rule-associated candidate node pair, based on the number of subsequent candidate nodes affected by the state change of the second candidate node in each second candidate node pair, determine the state transmission influence of each second candidate node, and determine the second candidate node with a state transmission influence greater than a preset influence threshold as a potential candidate node. According to the matching of the state change of each potential candidate node in the backtracking path with the business rules in the rule topology network, determine the potential candidate node that first appears with a state change inconsistent with the business rules, and determine the earliest inconsistent potential candidate node as the potential starting node. The first candidate node and the second candidate node are the previous nodes in the candidate node pair.

6. The knowledge graph and rule center-based metrology material accounting exception identification method of claim 5, wherein, The specific steps of determining the causal rule-associated candidate node pair include: Based on the state difference information related to the rule conflict of the first candidate node in the first candidate node pair, determine whether the first candidate node satisfies the condition of triggering the state change related to the rule conflict in the subsequent candidate node. If the state change condition is met, a logic mode of rule conflict propagation is determined based on a relationship between a state change of the first candidate node in the first candidate node pair and the business rule; Based on the logic mode of the first candidate node, it is determined whether there is a reasonable logic path of rule conflict state change from the state change of the first candidate node to the subsequent candidate node; If there is a reasonable logic path, the first candidate node pair with the reasonable logic path is determined as the cause-effect rule association candidate node pair.

7. The knowledge graph and rule center-based metrology material accounting exception identification method according to any one of claims 1 to 6, characterized in that, The abnormal accounting reason of the measured material is determined based on the abnormal operation mode of the potential starting node, including: The abnormal operation mode is split into multiple independent accounting process links, and a time sequence relationship between each accounting process link is obtained based on the time sequence between each accounting process link; Based on the time sequence relationship, associated accounting actions between adjacent accounting process links are determined, and each associated accounting action is analyzed for action specification based on a preset accounting specification action to obtain an initial abnormal accounting action that does not conform to the specification; Based on the mutual influence relationship between the abnormal accounting process links corresponding to the initial abnormal accounting action, an abnormal link influence relationship network is constructed, and a target abnormal accounting action that triggers subsequent accounting problems is determined based on the abnormal link influence relationship network combined with the business logic of the accounting process of the measured material. The abnormal accounting reason is determined based on the target abnormal accounting action.

8. A knowledge graph and rule center-based measurement material accounting exception identification system, characterized in that, The system is applied to the measured material accounting abnormality identification method based on the knowledge graph and the rule center in any one of claims 1 to 7; the system comprises: An accounting node analysis module is configured to determine state change information of the measured material at each accounting stage based on an entity state transition track of an accounting process of the measured material in a pre-constructed knowledge graph, and determine target accounting nodes in the accounting process based on the state change information at each accounting stage. A rule conflict analysis module is configured to perform rule conflict analysis based on node state information of each target accounting node and its associated business rules in a preset rule topology network to obtain abnormal accounting nodes. An accounting abnormality analysis module is configured to backtrack a state transition path before the abnormal accounting node based on the entity state transition track to obtain candidate accounting nodes in the backtracking process, and perform accounting abnormality analysis based on node states between the abnormal accounting node and each candidate accounting node and between each candidate accounting node to determine a potential starting node that causes rule conflict of the abnormal accounting node. An accounting abnormality cause analysis module is configured to determine an abnormal accounting reason of the measured material based on an abnormal operation mode of the potential starting node.

9. An electronic device comprising: A memory is configured to store a computer software program; A processor is configured to read and execute the computer software program, wherein the processor, when executing the computer software program, implements the measured material accounting abnormality identification method based on the knowledge graph and the rule center in any one of claims 1 to 7.

10. A non-transitory computer readable storage medium having stored therein a computer software program, characterized in that, The computer software program, when executed by a processor, implements the knowledge graph and rule center-based metered material accounting exception identification method according to any one of claims 1 to 7.

Images