Reimbursement data analysis and verification method, device and medium

Abstract

The present invention provides a reimbursement data analysis and verification method, device, and medium, which belongs to the field of enterprise reimbursement data processing technology. The method obtains the reimbursement associated fields of new financial accounting data and determines the correlation value between the reimbursement associated fields and the historical associated fields in the financial data set; obtains the historical financial accounting data corresponding to the historical associated fields whose correlation value is greater than the matching threshold, and uses the obtained historical financial accounting data as the comparison financial accounting data; and determines the reimbursement form verification result of the new financial accounting data based on the comparison financial accounting data and the new financial accounting data. The method comprehensively and quantitatively analyzes the correlation and similarity between new and old data from obtaining reimbursement associated fields, screening comparison data, to in-depth comparison based on financial semantic structure diagrams. While improving verification efficiency, it enhances accuracy and reliability, effectively avoiding the inefficiency and error-prone problems of traditional manual verification.

Description

Technical Field

[0001] The present invention belongs to the technical field of enterprise reimbursement data processing, and in particular relates to a reimbursement data analysis and verification method, device and medium. Background Art

[0002] Verifying financial accounting data is a fundamental and critical task in corporate financial management and expense reimbursement processes. As companies expand and their business complexity increases, financial systems must process a large volume of expense reimbursement documents daily. These data come from a wide range of sources, formats, and structures, posing significant challenges to verification.

[0003] Conventional technology requires manual or system-based verification of expense reports to search through all historical financial data for comparables, which is time-consuming and labor-intensive when faced with massive amounts of data. For example, large enterprises process thousands of expense reports daily. Comparing each report individually against historical data can lead to review cycles lasting several days, impacting capital flow efficiency.

[0004] Some companies use office software to assist with verification, but they rely solely on things like amount matching and invoice number verification, easily overlooking the underlying logical relationships between data. For example, it's difficult to detect issues like expense items on reimbursement forms that don't match the actual business scenario or data format errors, which can lead to false reimbursements or incorrect payments, causing economic losses to the company. Existing technologies make it difficult to dynamically match audit rules for different business scenarios, such as travel expenses and procurement expenses. For example, the review of travel expenses requires attention to the rationality of the itinerary, while procurement expenses require verification of supplier qualifications. Unified rules cannot meet diverse needs, resulting in some abnormal reimbursement forms being missed. Summary of the Invention

[0005] The present invention provides a reimbursement data analysis and verification method, which improves audit efficiency compared with manual audits in related technologies; through multi-dimensional quantitative evaluation, it effectively avoids problems such as false reimbursements and data errors, and ensures the authenticity and compliance of corporate financial data.

[0006] Methods include:

[0007] S101: Acquire a reimbursement-related field of new financial accounting data, and determine a correlation value between the reimbursement-related field and a historical correlation field in a financial data set;

[0008] S102: Acquire historical financial accounting data corresponding to historical correlation fields whose correlation values ​​are greater than a matching threshold, and use the acquired historical financial accounting data as comparison financial accounting data;

[0009] S103: Determine a reimbursement form verification result of the new financial accounting data based on the compared financial accounting data and the new financial accounting data.

[0010] Preferably, determining the correlation value between the reimbursement correlation field and the historical correlation field in the financial data set in step S101 specifically includes:

[0011] Obtain the reimbursement-related fields of the new financial accounting data and analyze their matching degree with the historical-related fields;

[0012] Count all matching results and determine the highest and lowest matching degrees as the basis for calculating the correlation value;

[0013] Combine the highest matching degree, the lowest matching degree, and the matching degree of each historical field to generate the corresponding correlation value;

[0014] According to the numerical sorting of correlation, historical field data with a value above the threshold is selected as comparison data for verifying the reimbursement form of the new financial accounting data.

[0015] Preferably, the method further obtains the difference between the highest matching degree and the lowest matching degree as a global range parameter;

[0016] For each historical association field, calculate the difference between its matching degree with the reimbursement association field and the minimum matching degree as the local offset;

[0017] Divide the local offset by the global range parameter to get the relative ratio as the correlation value;

[0018] Filter and compare financial accounting data based on correlation values.

[0019] Preferably, in step S101, obtaining the reimbursement-related fields of the new financial accounting data specifically includes:

[0020] Clean up invalid records of new financial accounting data to obtain preliminary processed data;

[0021] Match data format specifications according to reimbursement type, verify and correct data format;

[0022] Filter non-business content, retain core business data, and generate pre-processed expense reports;

[0023] The pre-processed data is encoded using a structured information parsing tool to generate reimbursement-related fields with fixed length and clear semantics.

[0024] Preferably, the method further comprises:

[0025] Multiple classification storage units are constructed by bill code location, and each classification storage unit only contains historical financial accounting data corresponding to a specific bill code;

[0026] Get the new bill code of the new financial accounting data;

[0027] Traverse all classified storage units, match the new bill code with the preset historical bill code in the classified storage unit, and filter out the target storage unit;

[0028] Extracting historical associated fields and their corresponding historical financial accounting data from the target storage unit;

[0029] Based on the semantic and grammatical analysis results of the reimbursement-related fields and the historical-related fields, the correlation value is calculated to complete the reimbursement form verification.

[0030] Preferably, step S103 further includes:

[0031] Obtaining the financial semantic structure diagram of new financial accounting data and comparing it with the historical financial semantic template of the financial accounting data;

[0032] Determine the reimbursement form verification results of the new financial accounting data based on the financial semantic structure diagram and historical financial semantic templates.

[0033] Preferably, step S103 further includes:

[0034] By analyzing the bill code prefix, keyword matching, and form title of new financial accounting data, it is classified into the preset business scenario classification system;

[0035] Identify the core business entities and their relationships within the classified new financial data;

[0036] Map the extracted entities to the predefined financial semantic metamodel, unify the entity expression form, standardize the attribute value format, and construct a standardized financial semantic structure diagram;

[0037] Based on the business scenario classification of new financial data, a collection of historical templates of the same type is retrieved from the historical financial semantic template library to form a comparison sample pool;

[0038] Compare the topological structures of the newly constructed financial semantic structure graph with those of the historical template, count the matching ratios of the same entity types and relationship paths, and generate a structural similarity score;

[0039] For the matched entity nodes, verify the consistency of their attribute values ​​and mark the abnormal level for the nodes with differences;

[0040] Combining the structural similarity score and attribute consistency verification results, a multi-dimensional comparison report is generated. When the similarity exceeds the preset threshold and there are no abnormalities in the key attributes, the reimbursement form is determined to have passed the verification.

[0041] Preferably, step S102 further extracts a financial semantic structure graph of the new financial accounting data; the financial semantic structure graph displays the relationship between financial data in the form of nodes and edges, wherein nodes represent different financial entities and edges represent the association between entities;

[0042] Retrieve historical financial semantic templates; these templates store structured information about past financial data in a standardized format, also presented as nodes and edges, and are verified to comply with established financial rules and reimbursement policies.

[0043] Through the semantic comparison algorithm, the financial semantic structure diagram of the new financial accounting data is compared item by item with the historical financial semantic template;

[0044] During the comparison process, pay attention to the consistency and differences of node types, node attributes, edge connections, and data flow paths;

[0045] Calculate the matching degree between the two according to the preset matching rules and matching threshold;

[0046] When the matching degree reaches a certain threshold, the new data is considered to be highly matched with certain historical data. Based on this, combined with the reimbursement rules, the compliance and accuracy of the new data are judged to obtain the reimbursement form verification result.

[0047] According to another embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the reimbursement data analysis and verification method when executing the program.

[0048] According to another embodiment of the present application, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the reimbursement data analysis and verification method are implemented.

[0049] It can be seen from the above technical solutions that the present invention has the following advantages:

[0050] The reimbursement data analysis and verification method provided by this invention establishes a quantitative correlation between new financial accounting data and historical data by extracting reimbursement-related fields and calculating correlation values. This method accurately selects highly correlated historical data as comparison samples, avoiding blind searches across all historical data and significantly improving data retrieval efficiency. By filtering historical financial accounting data based on correlation thresholds, the data scope is further narrowed, focusing on the most valuable comparison data and improving the specificity of the verification.

[0051] It also classifies business scenarios based on bill codes, keywords, and form titles to quickly locate the business type to which new data belongs, laying the foundation for the precise application of subsequent audit rules. It avoids confusion of data from different business scenarios and reduces the problem of mismatching audit rules. It identifies core business entities and relationships, and achieves standardization through semantic metamodel mapping to eliminate differences in data representation and format. A unified data representation facilitates cross-data comparison and improves data consistency and standardization. It retrieves and compares historical templates of the same type, and uses historical compliance data to quickly determine the rationality of new data structures and promptly identify abnormal business structures. It also converts new and historical data into visual semantic structure diagrams and templates, clearly presenting data logic in a node-edge format. The semantic comparison algorithm compares each item from multiple dimensions, such as nodes, edges, and data flows, and combines matching rules with quantitative threshold judgments to ensure comprehensive comparisons while improving audit efficiency and accuracy, reducing the subjectivity and omissions of manual audits. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 a flow chart of the reconciliation methodology for reimbursement data analysis;

[0054] Figure 2 A flow chart of an embodiment of a method for analyzing and verifying reimbursement data;

[0055] Figure 3 A flow chart of another embodiment of a method for analyzing and verifying reimbursement data;

[0056] Figure 4 Schematic diagram of an electronic device. DETAILED DESCRIPTION

[0057] The following describes in detail the specific steps of the reimbursement data analysis and verification method involved in this application. Specific details such as specific system structures and technologies are provided for illustrative purposes, not for limitation, to facilitate a thorough understanding of the embodiments of this application. However, it should be apparent to those skilled in the art that this application may also be implemented in other embodiments without these specific details.

[0058] It should be understood that when used in this specification, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their collections. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0059] The phrases "one embodiment" or "some embodiments" described in this application mean that the particular features, structures, or characteristics described in the embodiment are included in one or more embodiments of the application. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. that appear in different places in this application do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized.

[0060] In embodiments of the present invention, computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including, but not limited to, object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (exemplarily, via the Internet using an Internet service provider).

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0062] See also Figure 1 The figure is a flow chart of a method for analyzing and verifying reimbursement data in a specific embodiment, the method comprising:

[0063] S101: Acquire a reimbursement-related field of new financial accounting data, and determine a correlation value between the reimbursement-related field and a historical correlation field in a financial data set.

[0064] In some embodiments, key information is extracted from the newly submitted financial accounting data to form reimbursement-related fields. These reimbursement-related fields include, but are not limited to, the name of the reimbursement recipient, the reimbursement date, the expense item, the amount, the invoice number, and so on. These newly extracted reimbursement-related fields are then compared one by one with the historically related fields stored in the financial dataset. During this comparison, each historically related field is analyzed for similarity or matching with the new field across various dimensions, ultimately yielding a quantitative correlation value that represents the closeness of the association between the new field and each historical field.

[0065] In some specific embodiments, the specific process further includes:

[0066] Step 1011: clean up invalid records of the new financial accounting data to obtain preliminarily processed data.

[0067] Step 1012: Match the data format specifications according to the reimbursement type, verify and correct the data format, such as unifying the currency unit and date format.

[0068] Step 1013: Filter non-business content, retain core business data, and generate a pre-processed reimbursement form.

[0069] Step 1014: Encode the pre-processed data using a structured information parsing tool to generate reimbursement-related fields with fixed length and clear semantics.

[0070] This embodiment uses a structured information parsing tool to encode the pre-processed expense report to generate fixed-length and semantically standardized expense-related fields. This process is based on predefined data format specifications to ensure the uniformity of expense-related fields in different financial accounting scenarios. For example, for a certain type of expense report, the tool will extract key information such as the amount, date, and supplier name, and convert it into a standardized code, such as a numeric identifier or hash value, to form a field of length L. This method solves the problem of field matching difficulties in traditional expense reports due to data heterogeneity by eliminating redundancy and format differences in the original data, thereby improving the efficiency and accuracy of subsequent verification.

[0071] By obtaining reimbursement-related fields and calculating correlation values, this embodiment can quickly locate historical data that may be associated with new financial accounting data, narrowing the scope of subsequent comparisons and improving data processing efficiency. Furthermore, the quantified correlation values ​​provide a clear basis for subsequent screening, avoiding the subjectivity and arbitrariness of human judgment.

[0072] S102: Acquire historical financial accounting data corresponding to historical correlation fields whose correlation values ​​are greater than a matching threshold, and use the acquired historical financial accounting data as comparison financial accounting data.

[0073] This embodiment sets a matching threshold to identify historically relevant fields with a sufficiently high correlation with the new financial accounting data. Based on the correlation value, historically relevant fields with correlations greater than the matching threshold are filtered from the financial dataset. The corresponding historical financial accounting data is then extracted and used as the comparison financial accounting data. By comparing the historically relevant fields with the matching threshold, the historical data with low correlations is automatically filtered out, while the data potentially highly correlated with the new data is retained as the basis for subsequent comparisons.

[0074] S103: Determine a reimbursement form verification result of the new financial accounting data based on the compared financial accounting data and the new financial accounting data.

[0075] The comparison financial accounting data obtained in step S102 is compared and analyzed with the new financial accounting data. This comparison covers various aspects of the data, such as the legitimacy of the reimbursement reason, the consistency of the expense amount, and the authenticity of the invoice information. The comparison results are judged and evaluated based on pre-set verification rules and standards to ultimately determine whether the reimbursement form based on the new financial accounting data meets the requirements. The verification result of the reimbursement form is then determined, such as passed, failed, or requiring further information.

[0076] For example, it checks whether the reimbursement amount is within the budget, whether the invoice is valid and not a duplicate, etc. Based on the inspection results, logical judgment is made in combination with rules and standards to reach the final verification conclusion. It can effectively identify problems and anomalies in new expense reports, ensuring their authenticity, legality, and compliance.

[0077] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, S101 further includes: constructing multiple classification storage units based on the bill code location, each classification storage unit only containing historical financial accounting data corresponding to a specific bill code;

[0078] Obtain the new invoice code for the new financial accounting data; match it to the corresponding classification storage unit and filter out historically associated field data that matches the invoice code. This method, through the classification storage design of invoice codes, avoids the inefficiency of traditional full-data retrieval and significantly improves the response speed of reimbursement form verification. Furthermore, the precise matching mechanism between the classification storage unit and the new invoice code ensures that the data scope of the correlation analysis focuses on historical data that aligns with the business logic.

[0079] It's important to note that historical financial accounting data is categorized by code characteristics using the location rules of bill codes (e.g., the first three digits represent the bill type, and the middle four digits represent the region where the bill was issued). For example, all accounting data corresponding to invoices with the first three digits "FP0" are stored in the same unit, forming a categorized storage unit for the "FP0 bill type." This enables structured data storage and improves data retrieval efficiency.

[0080] The system automatically extracts the invoice code field from newly submitted financial accounting data. For example, in electronic invoice data, the system identifies characters in fixed positions (such as digits 5-15 in the file header) or after a specific identifier to obtain a unique invoice code. Each classification storage unit is accessed sequentially, and the new invoice code is compared against the pre-set encoding rules within the unit. If the code matches exactly or meets a specific matching logic, meaning that the first digits are identical, the invoice is considered to be of the same type, the classification storage unit is marked as the target unit. After determining the target unit, the system reads all historical financial accounting data within that unit and extracts key related fields (such as the reimbursement recipient, amount, and expense item). These fields correspond to the reimbursement related field structure of the new data, facilitating subsequent comparisons. Related fields in the new and historical data undergo a dual analysis: semantic analysis determines whether the field meaning is consistent (for example, "travel expenses" and "business trip expenses" are considered the same type); and syntactic analysis checks whether the field format matches (for example, whether the amount is in numeric format and whether the date conforms to a standard format). Based on the analysis results, the correlation between the fields is comprehensively assessed to quantify the correlation value. This dual semantic and syntactic verification improves verification accuracy.

[0081] In an embodiment of the present invention, based on the determination of the correlation value between the reimbursement correlation field and the historical correlation field in the financial data set in step S101, a possible embodiment will be given below to illustrate its specific implementation scheme in a non-limiting manner.

[0082] Obtain the reimbursement-related fields of the new financial accounting data and analyze their matching with the historically related fields. Count all matching results, determine the highest and lowest matching scores, and use them as the basis for calculating the correlation value. Combine the highest and lowest matching scores with the matching scores of each historical field to generate the corresponding correlation value. Sort by correlation value and select historical field data that exceeds the threshold as comparison data for verifying the reimbursement forms of the new financial accounting data.

[0083] The method of this embodiment also obtains the difference between the highest matching degree and the lowest matching degree as a global scope parameter; for each historical association field, calculates the difference between its matching degree with the reimbursement association field and the lowest matching degree as a local offset; divides the local offset by the global scope parameter to obtain a relative proportion as a correlation value; and filters and compares financial accounting data based on the correlation value.

[0084] It should be noted that the global range parameter represents the distribution range of the matching scores of all historically associated fields by calculating the difference between the highest and lowest matching scores. This parameter is used to quantify the overall fluctuation range of historical data. The local offset calculates the difference between the matching score of the reimbursement-related field and its lowest matching score for each historically associated field, representing the relative position of the current field within the overall range. The relative ratio is calculated by dividing the local offset by the global range parameter to obtain the relative ratio between the reimbursement-related field and the historically associated field. This ratio reflects the normalized position of the current field's matching score within the overall distribution, thereby generating the correlation value.

[0085] As can be seen, reimbursement-related fields, such as the claimant, reimbursement amount, expense item, and invoice number, are extracted from the newly submitted financial accounting data. Next, these new fields are compared one by one with the related fields in the historical financial data, analyzing the degree of match between the two and determining the matching score for each historical field. Subsequently, all matching scores are statistically analyzed to identify the highest and lowest matching scores, which serve as the basis for subsequent correlation calculations. A specific calculation method is then used to generate a corresponding correlation score, combining the highest and lowest matching scores with the matching score of each historical field itself. This score quantifies the closeness of the correlation between the new field and each historical field. Finally, the correlation scores of all historical fields are sorted, and a threshold is set. Historical fields with correlation scores exceeding the threshold, along with their corresponding financial accounting data, are selected as the comparison data for verification against the new financial accounting data, completing the preliminary data screening for reimbursement form verification.

[0086] For example, fields such as "Reimburser Name," "Amount Numerical," and "Invoice Number String" are extracted from electronic expense reports. Then, for each historically associated field, different matching rules are set based on the field type: for text fields (such as the reimburser's name), a character similarity algorithm is used to determine if they are identical or similar; for numeric fields (such as the amount), a tolerance range (such as ±5%) is set to determine the degree of numerical similarity; and for numbered fields (such as the invoice number), an exact match is required. Using these rules, the new field is compared against the historical fields one by one to determine the matching degree for each historical field.

[0087] Of course, this embodiment can also use the formula "(matching degree of a certain historical field - lowest matching degree) / (highest matching degree - lowest matching degree)" to calculate the relative ratio, and use this ratio as the correlation value. This process maps the original matching degree to a range of 0-1, with the value closer to 1 indicating a closer association between the historical field and the new field. Converting matching results of different magnitudes and dimensions into a unified quantitative indicator eliminates interference caused by data differences and improves data processing efficiency.

[0088] This embodiment also first calculates the difference between the highest and lowest matching degrees. This difference reflects the overall fluctuation range of the matching degrees of all historical fields and serves as a global range parameter. Then, for each historically associated field, the difference between its matching degree and the lowest matching degree is calculated to obtain a local offset, which represents the degree of deviation of the field from the overall matching level. Finally, the local offset is divided by the global range parameter to obtain a relative ratio. This ratio is the correlation value, which is used to measure the strength of the association between the historical field and the new field. This prevents a single extreme value from affecting the overall judgment and further improves the accuracy of the screening and comparison data.

[0089] In one embodiment of the present application, in order to obtain a more accurate reimbursement form verification result, the present application adopts a reimbursement form verification mechanism, and step S103 further includes:

[0090] Obtain a financial semantic structure diagram of the new financial accounting data and compare it with a historical financial semantic template of the financial accounting data. Determine a reimbursement form verification result of the new financial accounting data based on the financial semantic structure diagram and the historical financial semantic template.

[0091] It's important to note that the financial semantic structure graph is a structured semantic network, with nodes representing business entities (e.g., "reimbursement recipient," "amount") and edges representing relationships between entities (e.g., "belongs to," "associated with"). Working process: Semantic analysis is used to identify key entities and relationships in expense reports, constructing a directed graph structure. Each node is labeled with the entity type (e.g., "person," "expense item") and attribute value (e.g., "Zhang San," "500 yuan").

[0092] The historical financial semantic template performs cluster analysis on historical expense reports, extracts high-frequency entities and relationship patterns, and forms a standardized template library. Each template represents a typical reimbursement business (such as "travel expense reimbursement" and "office supply procurement").

[0093] Business scenario classification is based on analyzing the prefix of the bill code (such as "TRAVEL-" for travel expenses) or keywords (such as "air tickets" and "accommodation fees"), and classifying the reimbursement form into the preset business scenarios (such as "travel", "procurement", and "meeting").

[0094] Business entity normalization involves mapping synonymous entities to standard terms (such as "computer," "electronic equipment"), normalizing numeric formats (such as "￥5,000," "5000.00"), and removing formatting symbols (such as "Invoice Number: FP20250525" → "FP20250525") to ensure consistency in entity expression.

[0095] In this embodiment, business scenario identification analyzes invoice code prefixes or keywords to determine the business scenario (e.g., "travel expenses," "office supplies procurement") to which new financial accounting data belongs. Semantic structure construction, based on the identified business scenario, extracts entities (e.g., "reimburser," "amount," "supplier") and their relationships within the new financial data to construct a financial semantic structure graph. Template matching selects historical templates from a library of historical financial semantic templates that match the current business scenario, forming a comparison sample set.

[0096] In this embodiment, the business scenario classification of the new financial accounting data is determined based on the bill code of the new financial accounting data. Based on the business scenario classification, a target parsing library is determined. The new financial accounting data and the comparison financial accounting data are parsed based on the target parsing library to obtain a financial semantic structure diagram for the new financial accounting data and a historical financial semantic template for the comparison financial accounting data. When parsing the new and comparison financial accounting data, the target parsing library requires business entity normalization to ensure structural consistency within the semantics.

[0097] This embodiment extracts the type (such as "personnel" and "expense item") and the normalized attribute value (such as "Zhang San" and "500.00") for each node in the financial semantic structure diagram and the history template.

[0098] The node type and attribute value are combined to generate a hash value, which serves as the node's unique identifier. The node hash values ​​of the new financial data are compared with those of the historical template, and the proportion of identical nodes is calculated as the structural match. The historical template with the highest match is selected as the comparison result. If the match exceeds a threshold (e.g., 95%), the reimbursement form is considered compliant.

[0099] On the basis of the above embodiment, in order to further improve the reliability of the reimbursement data analysis and verification method provided in the above embodiment, the following is a more specific implementation method, which includes:

[0100] S901, pre-processing the new financial accounting data to obtain pre-processed financial accounting data.

[0101] S902: Encode the pre-processed financial accounting data based on a structured information parsing tool to obtain reimbursement-related fields of new financial accounting data.

[0102] S903, obtaining a new bill code for the new financial accounting data.

[0103] S904: Select a financial data set from the classification storage unit based on the new invoice code and the historical invoice codes of the historical financial accounting data corresponding to each historical association field in the classification storage unit. The historical invoice codes of the historical financial accounting data corresponding to each historical association field in each classification storage unit are identical; and the historical invoice codes of the historical financial accounting data corresponding to each historical association field in the financial data set are identical to the new invoice code.

[0104] S905: Determine the matching degree between the reimbursement related fields and the historical related fields in the financial data set.

[0105] S906: Select the highest matching degree and the lowest matching degree from the matching degrees.

[0106] S907: The difference between the highest matching degree and the lowest matching degree is used as a global range parameter.

[0107] S908: For each history-related field, calculate the difference between its matching degree with the reimbursement-related field and the minimum matching degree, and use it as a local offset.

[0108] S909: Divide the local offset by the global range parameter to obtain a relative ratio as a correlation value.

[0109] S910 , obtaining historical financial accounting data corresponding to historical correlation fields whose correlation values ​​are greater than a matching threshold, and using the obtained historical financial accounting data as comparison financial accounting data.

[0110] S911, obtaining a financial semantic structure diagram of new financial accounting data and comparing it with a historical financial semantic template of the financial accounting data.

[0111] S912: Determine the reimbursement form verification result of the new financial accounting data based on the financial semantic structure diagram and the historical financial semantic template.

[0112] like Figure 2 As shown, as step S103 of this application, in addition to the above method, another implementation method is also included as follows:

[0113] Step S1031: Analyze the bill code prefix, keyword matching, and form title of the new financial accounting data and classify it into a preset business scenario classification system;

[0114] Step S1032: identifying the core business entities and their relationships within the classified new financial data;

[0115] Step S1033: Mapping the extracted entities with the predefined financial semantic metamodel, unifying the entity expression form, and normalizing the attribute value format to construct a standardized financial semantic structure diagram;

[0116] Step S1034: Based on the business scenario classification of the new financial data, a historical template set of the same type is retrieved from the historical financial semantic template library to form a comparison sample pool;

[0117] Step S1035: Compare the topological structures of the newly constructed financial semantic structure graph with those of the historical template, count the matching ratios of the same entity types and relationship paths, and generate a structural similarity score;

[0118] Step S1036: Verify the consistency of attribute values ​​of the matched entity nodes and mark abnormal levels for nodes with differences;

[0119] Step S1037: Combine the structural similarity score and the attribute consistency verification results to generate a multi-dimensional comparison report. When the similarity exceeds the preset threshold and there is no abnormality in the key attributes, the reimbursement form is determined to have passed the verification.

[0120] As can be seen, by analyzing the bill code prefixes, keyword matching, and form titles of the new financial accounting data, it is classified into a pre-defined business scenario classification system, laying the foundation for subsequent processing. For the classified data, core business entities and their associated relationships are identified, clarifying the key elements and interactions of the data. Next, the extracted entities are mapped to a pre-defined financial semantic metamodel, unifying the entity representation and normalizing the attribute value format to construct a standardized financial semantic structure diagram for structured comparison. Based on the business scenario classification of the new financial data, a collection of historical templates of the same type is retrieved from the historical financial semantic template library to form a comparison sample pool, providing historical data reference. The topology of the newly constructed financial semantic structure diagram is compared with that of the historical templates. The matching ratio of the same entity types and relationship paths is calculated, and a structural similarity score is generated to quantify the degree of similarity. For matched entity nodes, the consistency of their attribute values ​​is verified, and nodes with discrepancies are marked with anomaly levels to identify potential issues. Finally, a multi-dimensional comparison report is generated by combining the structural similarity score and attribute consistency verification results. When the similarity exceeds the preset threshold and there are no abnormalities in the key attributes, the reimbursement form is judged to have passed the verification and a clear conclusion is given.

[0121] This approach maps the extracted entities to a predefined financial semantic metamodel and normalizes the attribute value format, unifying data representation and eliminating comparison errors caused by inconsistent formats. A collection of similar historical templates from a historical financial semantic template library is retrieved to form a comparison sample pool, providing historical reference data that matches the new data's business scenarios, increasing the relevance and effectiveness of the comparison. Comparing the topology of the newly constructed financial semantic structure graph with the historical templates and generating a structural similarity score quantifies the degree of similarity between the new and old data, providing an objective basis for assessing the compliance and accuracy of expense claims. Verifying the consistency of attribute values ​​of matching entity nodes and marking discrepant nodes with anomaly levels accurately identifies specific differences between the new and old data, allowing for timely detection of potential errors or anomalies. Combining the structural similarity score and attribute consistency verification results to generate a multi-dimensional comparison report comprehensively presents the comparison of the new and old data and provides clear verification conclusions. When the similarity and key attributes meet the requirements, the expense claim is considered approved, effectively improving the automation of expense claim processing and the accuracy of decision-making.

[0122] like Figure 3 As shown, as a method of this application, step S102 also includes the following implementation method:

[0123] Step S4011: extracting a financial semantic structure graph of the new financial accounting data; the financial semantic structure graph displays the relationship between financial data in the form of nodes and edges, where nodes represent different financial entities and edges represent the association between entities;

[0124] Step S4012: Retrieve a historical financial semantic template; the historical financial semantic template stores structured information of past financial data in a standardized format, also presented in the form of nodes and edges, and has been verified to comply with established financial rules and reimbursement policies;

[0125] Step S4013: Using a semantic comparison algorithm, the financial semantic structure diagram of the new financial accounting data is compared item by item with the historical financial semantic template;

[0126] Step S4014: During the comparison process, focus on the consistency and differences of node types, node attributes, edge connections, and data flow paths;

[0127] Step S4015: Calculate the degree of match between the two based on the preset matching rules and matching threshold. When the matching degree reaches a certain threshold, the new data is considered to be highly matched with the historical data. Based on this, the compliance and accuracy of the new data are determined in combination with the reimbursement rules, thus obtaining the reimbursement form verification result.

[0128] For example, extracting a financial semantic structure graph from new financial accounting data involves in-depth analysis of the data, identifying key financial entities such as income, expenditure, and accounts as nodes, and the relationships between these entities, such as capital flow and ownership, as edges.

[0129] Step S4012 retrieves historical financial semantic templates from existing historical data stores. These templates are not only standardized in format but also align with the company's financial rules and reimbursement policies. They are also constructed using nodes and edges. Then, in step S4013, a specially designed semantic comparison algorithm is used to compare the semantic structure of the new data with the historical templates. This algorithm carefully examines node types, such as different types of income or expense categories, node attributes such as amount and time, and edge connections, such as which account funds flow from and to, as well as the consistency of the data flow paths.

[0130] In step S4014, the focus of the comparison work is to find out the similarities or differences between the new and old data in terms of node types, attributes, edge connection relationships and data flow paths. Finally, step S4015 calculates the degree of matching between the new and old data based on the pre-set matching rules and matching thresholds. If the matching degree reaches or exceeds the set threshold, it is considered that the new data is highly matched with a certain historical data. At this time, combined with the company's reimbursement rules, the compliance and accuracy of the new data can be judged, thereby obtaining the final result of the reimbursement form verification. The degree of matching is quantified by pre-set matching rules and thresholds, making the audit results more objective and repeatable. When the new data is highly matched with the historical data, the final judgment is made in combination with the reimbursement rules, which not only utilizes historical experience but also ensures compliance with current policy requirements, ensures that the audit results of the reimbursement form are reliable, and reduces the financial risks of the company.

[0131] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0132] like Figure 4 As shown, the present application also provides an electronic device, including a display module 103, a memory 102, a processor 101 and a computer program stored in the memory and executable on the processor 101, wherein the processor 101 implements the steps of the reimbursement data analysis and verification method when executing the program.

[0133] In the embodiments of the present invention, electronic devices include, but are not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or claimed herein.

[0134] In the embodiment of the present application, the processor 101 can be implemented by using at least one of a special purpose integrated circuit, a programmable logic device, a field programmable gate array, a processor, a controller, a microcontroller, a microprocessor, and an electronic unit designed to perform the functions described herein. In some cases, such an embodiment can be implemented in a controller. For software implementation, an embodiment such as a process or function can be implemented with a separate software module that allows the execution of at least one function or operation. The software code can be implemented by a software application (or program) written in any appropriate programming language, and the software code can be stored in a memory and executed by a controller.

[0135] The display module 103 is used to display information input by the user or information provided to the user. The display module 103 may include a display panel, which may be configured in the form of a liquid crystal display, an organic light emitting diode, etc.

[0136] The memory 102 can be used to store software programs and various data. The memory 102 can include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0137] The present application also provides a storage medium having a computer program stored thereon, which implements the steps of the reimbursement data analysis and verification method when executed by a processor.

[0138] The storage medium can be any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0139] In the context of storage media, a readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0140] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for analyzing and checking reimbursement data, characterized in that: include: S101: Acquire a reimbursement-related field of new financial accounting data, and determine a correlation value between the reimbursement-related field and a historical correlation field in a financial data set; Specifically include: Obtain the reimbursement-related fields of the new financial accounting data and analyze their matching degree with the historical-related fields; Count all matching results and determine the highest and lowest matching degrees as the basis for calculating the correlation value; Combine the highest matching degree, the lowest matching degree, and the matching degree of each historical field to generate the corresponding correlation value; Sort by correlation value and select historical field data above the threshold as comparison data to verify the reimbursement form of the new financial accounting data; S102: Acquire historical financial accounting data corresponding to historical correlation fields whose correlation values ​​are greater than a matching threshold, and use the acquired historical financial accounting data as comparison financial accounting data; S103: Determining a reimbursement form verification result of the new financial accounting data based on the compared financial accounting data and the new financial accounting data; Obtaining the financial semantic structure diagram of new financial accounting data and comparing it with the historical financial semantic template of the financial accounting data; Determine the reimbursement form verification results of the new financial accounting data based on the financial semantic structure diagram and historical financial semantic templates.

2. The reimbursement data analysis and verification method according to claim 1, characterized in that: The method also obtains the difference between the highest matching degree and the lowest matching degree as a global scope parameter; For each historical association field, calculate the difference between its matching degree with the reimbursement association field and the minimum matching degree as the local offset; Divide the local offset by the global range parameter to get the relative ratio as the correlation value; Filter and compare financial accounting data based on correlation values.

3. The reimbursement data analysis and verification method according to claim 1, characterized in that: In step S101, obtaining the reimbursement-related fields of the new financial accounting data specifically includes: Clean up invalid records of new financial accounting data to obtain preliminary processed data; Match data format specifications according to reimbursement type, verify and correct data format; Filter non-business content, retain core business data, and generate pre-processed expense reports; The pre-processed data is encoded using a structured information parsing tool to generate reimbursement-related fields with fixed length and clear semantics.

4. The reimbursement data analysis and verification method according to claim 1, characterized in that: The method also includes: Multiple classification storage units are constructed by bill code location, and each classification storage unit only contains historical financial accounting data corresponding to a specific bill code; Get the new bill code of the new financial accounting data; Traverse all classified storage units, match the new bill code with the preset historical bill code in the classified storage unit, and filter out the target storage unit; Extracting historical associated fields and their corresponding historical financial accounting data from the target storage unit; Based on the semantic and grammatical analysis results of the reimbursement-related fields and the historical-related fields, the correlation value is calculated to complete the reimbursement form verification.

5. The reimbursement data analysis and verification method according to claim 1, characterized in that: Step S103 further includes: By analyzing the bill code prefix, keyword matching, and form title of new financial accounting data, it is classified into the preset business scenario classification system; Identify the core business entities and their relationships within the classified new financial data; Map the extracted entities to the predefined financial semantic metamodel, unify the entity expression form, standardize the attribute value format, and construct a standardized financial semantic structure diagram; Based on the business scenario classification of new financial data, a collection of historical templates of the same type is retrieved from the historical financial semantic template library to form a comparison sample pool; Compare the topological structures of the newly constructed financial semantic structure graph with those of the historical template, count the matching ratios of the same entity types and relationship paths, and generate a structural similarity score; For the matched entity nodes, verify the consistency of their attribute values ​​and mark the abnormal level for the nodes with differences; Combining the structural similarity score and attribute consistency verification results, a multi-dimensional comparison report is generated. When the similarity exceeds the preset threshold and there are no abnormalities in the key attributes, the reimbursement form is determined to have passed the verification.

6. The reimbursement data analysis and verification method according to claim 1, characterized in that: Step S102 also extracts a financial semantic structure graph of the new financial accounting data; the financial semantic structure graph displays the relationship between financial data in the form of nodes and edges, where nodes represent different financial entities and edges represent the relationship between entities; Retrieve historical financial semantic templates; these templates store structured information about past financial data in a standardized format, also presented as nodes and edges, and are verified to comply with established financial rules and reimbursement policies. Through the semantic comparison algorithm, the financial semantic structure diagram of the new financial accounting data is compared item by item with the historical financial semantic template; During the comparison process, pay attention to the consistency and differences of node types, node attributes, edge connections, and data flow paths; Calculate the matching degree between the two according to the preset matching rules and matching threshold; When the matching degree reaches a certain threshold, the new data is considered to be highly matched with certain historical data. Based on this, combined with the reimbursement rules, the compliance and accuracy of the new data are judged to obtain the reimbursement form verification result.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the reimbursement data analysis and verification method according to any one of claims 1 to 6 are implemented.

8. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the reimbursement data analysis and verification method according to any one of claims 1 to 6 are implemented.

Images